Search results for: hearing reconstruction

Authors: H. N. Alyaseen, S. E. Alalawi, T. Cordoba, É. Delisle, C. Cordoba, A. Odobescu

Abstract:

Sternal dehiscence is defined as the persistent separation of sternal bones that are often complicated with mediastinitis. Etiologies that lead to sternal dehiscence vary, with cardiovascular and thoracic surgeries being the most common. Early diagnosis in susceptible patients is crucial to the management of such cases, as they are associated with high mortality rates. A recent meta-analysis of more than four hundred thousand patients concluded that deep sternal wound infections were the leading cause of mortality and morbidity in patients undergoing cardiac procedures. Long-term complications associated with sternal dehiscence include increased hospitalizations, cardiac infarctions, and renal and respiratory failures. Numerous osteosynthesis methods have been described in the literature. Surgical materials offer enough rigidity to support the sternum and can be flexible enough to allow physiological breathing movements of the chest; however, these materials fall short when managing patients with extensive bone loss, osteopenia, or general poor bone quality, for such cases, flaps offer a better closure system. Early utilization of flaps yields better survival rates compared to delayed closure or to patients treated with sternal rewiring and closed drainage. The utilization of pectoralis major flaps, rectus abdominus, and latissimus muscle flaps have all been described in the literature as great alternatives. Flap selection depends on a variety of factors, mainly the size of the sternal defect, infection, and the availability of local tissues. Free fibular flaps are commonly harvested flaps utilized in reconstruction around the body. In cases regarding sternal reconstruction with free fibular flaps, the literature exclusively discussed the flap applied vertically to the chest wall. We present a different technique applying the free fibular triple barrel flap oriented in a transverse manner, in parallel to the ribs. In our experience, this method could have enhanced results and improved prognosis as it contributes to the normal circumferential shape of the chest wall.

Keywords: sternal dehiscence, management, free fibular flaps, novel surgical techniques

Procedia PDF Downloads 96

Authors: Altayeb Abdalla Ahmed

Abstract:

Skeletal phenotype is a product of a balanced interaction between genetics and environmental factors throughout different life stages. Therefore, interlimb proportions are variable between populations. Although interlimb proportion indices have been used in anthropology in assessing the influence of various environmental factors on limbs, an extensive literature review revealed that there is a paucity of published research assessing interlimb part correlations and possibility of reconstruction. Hence, this study aims to assess the relationships between upper and lower limb parts and develop regression formulae to reconstruct the parts from one another. The left upper arm length, ulnar length, wrist breadth, hand length, hand breadth, tibial length, bimalleolar breadth, foot length, and foot breadth of 376 right-handed subjects, comprising 187 males and 189 females (aged 25-35 years), were measured. Initially, the data were analyzed using basic univariate analysis and independent t-tests; then sex-specific simple and multiple linear regression models were used to estimate upper limb parts from lower limb parts and vice-versa. The results of this study indicated significant sexual dimorphism for all variables. The results indicated a significant correlation between the upper and lower limbs parts (p < 0.01). Linear and multiple (stepwise) regression equations were developed to reconstruct the limb parts in the presence of a single or multiple dimension(s) from the other limb. Multiple stepwise regression equations generated better reconstructions than simple equations. These results are significant in forensics as it can aid in identification of multiple isolated limb parts particularly during mass disasters and criminal dismemberment. Although a DNA analysis is the most reliable tool for identification, its usage has multiple limitations in undeveloped countries, e.g., cost, facility availability, and trained personnel. Furthermore, it has important implication in plastic and orthopedic reconstructive surgeries. This study is the only reported study assessing the correlation and prediction capabilities between many of the upper and lower dimensions. The present study demonstrates a significant correlation between the interlimb parts in both sexes, which indicates a possibility to reconstruction using regression equations.

Keywords: anthropometry, correlation, limb, Sudanese

Procedia PDF Downloads 295

Authors: El Mahdi Barrah, Said Safi, Abdessamad Malaoui

Abstract:

In this paper, we present an algorithm for reconstruction from incomplete 2D scans for tifinagh characters. This algorithm is based on using correlation between the lost block and its neighbors. This system proposed contains three main parts: pre-processing, features extraction and recognition. In the first step, we construct a database of tifinagh characters. In the second step, we will apply “shape analysis algorithm”. In classification part, we will use Neural Network. The simulation results demonstrate that the proposed method give good results.

Keywords: Tifinagh character recognition, neural networks, local cost computation, ANN

Procedia PDF Downloads 335

Authors: J. Chamorro-Servent, S. Tassani, M. A. Gonzalez-Ballester, L. J. Roca, J. Romeu, O. Camara

Abstract:

Electromagnetic and microwave imaging (MWI) have been used in medical imaging in the last years, being the most common applications of breast cancer and stroke detection or monitoring. In those applications, the subject or zone to observe is surrounded by a number of antennas, and the Nyquist criterium can be satisfied. Additionally, the space between the antennas (transmitting and receiving the electromagnetic fields) and the zone to study can be prepared in a homogeneous scenario. However, this may differ in other cases as could be intracardiac catheters, stomach monitoring devices, pelvic organ systems, liver ablation monitoring devices, or uterine fibroids’ ablation systems. In this work, we analyzed different MWI algorithms to find the most suitable method for dealing with an intrabody scenario. Due to the space limitations usually confronted on those applications, the device would have a cylindrical configuration of a maximum of eight transmitters and eight receiver antennas. This together with the positioning of the supposed device inside a body tract impose additional constraints in order to choose a reconstruction method; for instance, it inhabitants the use of well-known algorithms such as filtered backpropagation for diffraction tomography (due to the unusual configuration with probes enclosed by the imaging region). Finally, the difficulty of simulating a realistic non-homogeneous background inside the body (due to the incomplete knowledge of the dielectric properties of other tissues between the antennas’ position and the zone to observe), also prevents the use of Born and Rytov algorithms due to their limitations with a heterogeneous background. Instead, we decided to use a time-reversed algorithm (mostly used in geophysics) due to its characteristics of ignoring heterogeneities in the background medium, and of focusing its generated field onto the scatters. Therefore, a 2D time-reversed finite difference time domain was developed based on the time-reversed approach for microwave breast cancer detection. Simultaneously an in-silico testbed was also developed to compare ground-truth dielectric properties with corresponding microwave imaging reconstruction. Forward and inverse problems were computed varying: the frequency used related to a small zone to observe (7, 7.5 and 8 GHz); a small polyp diameter (5, 7 and 10 mm); two polyp positions with respect to the closest antenna (aligned or disaligned); and the (transmitters-to-receivers) antenna combination used for the reconstruction (1-1, 8-1, 8-8 or 8-3). Results indicate that when using the existent time-reversed method for breast cancer here for the different combinations of transmitters and receivers, we found false positives due to the high degrees of freedom and unusual configuration (and the possible violation of Nyquist criterium). Those false positives founded in 8-1 and 8-8 combinations, highly reduced with the 1-1 and 8-3 combination, being the 8-3 configuration de most suitable (three neighboring receivers at each time). The 8-3 configuration creates a region-of-interest reduced problem, decreasing the ill-posedness of the inverse problem. To conclude, the proposed algorithm solves the main limitations of the described intrabody application, successfully detecting the angular position of targets inside the body tract.

Keywords: FDTD, time-reversed, medical imaging, microwave imaging

Procedia PDF Downloads 127

Authors: Anjalika Fernando, Banuka Athuraliya

Abstract:

This paper presents a comprehensive approach for converting Sinhala Sign Language (SSL) into grammatically correct sentences using Natural Language Processing (NLP) techniques in real-time. While previous studies have explored various aspects of SSL translation, the research gap lies in the absence of grammar checking for SSL. This work aims to bridge this gap by proposing a two-stage methodology that leverages deep learning models to detect signs and translate them into coherent sentences, ensuring grammatical accuracy. The first stage of the approach involves the utilization of a Long Short-Term Memory (LSTM) deep learning model to recognize and interpret SSL signs. By training the LSTM model on a dataset of SSL gestures, it learns to accurately classify and translate these signs into textual representations. The LSTM model achieves a commendable accuracy rate of 94%, demonstrating its effectiveness in accurately recognizing and translating SSL gestures. Building upon the successful recognition and translation of SSL signs, the second stage of the methodology focuses on improving the grammatical correctness of the translated sentences. The project employs a Neural Machine Translation (NMT) architecture, consisting of an encoder and decoder with LSTM components, to enhance the syntactical structure of the generated sentences. By training the NMT model on a parallel corpus of Sinhala wrong sentences and their corresponding grammatically correct translations, it learns to generate coherent and grammatically accurate sentences. The NMT model achieves an impressive accuracy rate of 98%, affirming its capability to produce linguistically sound translations. The proposed approach offers significant contributions to the field of SSL translation and grammar correction. Addressing the critical issue of grammar checking, it enhances the usability and reliability of SSL translation systems, facilitating effective communication between hearing-impaired and non-sign language users. Furthermore, the integration of deep learning techniques, such as LSTM and NMT, ensures the accuracy and robustness of the translation process. This research holds great potential for practical applications, including educational platforms, accessibility tools, and communication aids for the hearing-impaired. Furthermore, it lays the foundation for future advancements in SSL translation systems, fostering inclusive and equal opportunities for the deaf community. Future work includes expanding the existing datasets to further improve the accuracy and generalization of the SSL translation system. Additionally, the development of a dedicated mobile application would enhance the accessibility and convenience of SSL translation on handheld devices. Furthermore, efforts will be made to enhance the current application for educational purposes, enabling individuals to learn and practice SSL more effectively. Another area of future exploration involves enabling two-way communication, allowing seamless interaction between sign-language users and non-sign-language users.In conclusion, this paper presents a novel approach for converting Sinhala Sign Language gestures into grammatically correct sentences using NLP techniques in real time. The two-stage methodology, comprising an LSTM model for sign detection and translation and an NMT model for grammar correction, achieves high accuracy rates of 94% and 98%, respectively. By addressing the lack of grammar checking in existing SSL translation research, this work contributes significantly to the development of more accurate and reliable SSL translation systems, thereby fostering effective communication and inclusivity for the hearing-impaired community

Keywords: Sinhala sign language, sign Language, NLP, LSTM, NMT

Procedia PDF Downloads 106

Authors: Anastasia Poliakova

Abstract:

Harmful algal bloom (HAB) is a known negative phenomenon that is caused both by natural factors and anthropogenic influence. HABs can result in a series of deleterious effects, such as beach fouling, paralytic shellfish poisoning, mass mortality of marine species, and a threat to human health, especially if toxins pollute drinking water or occur nearby public resorts. In South China, the problem of HABs has an ultimately important meaning. For this study, we used a 1.5 m sediment core LAX-2018-2 collected in 2018 from the Zhanjiang Mangrove National Nature Reserve (109°03´E, 20°30´N), Guangdong Province, South China. High-resolution coastal environment reconstruction with a specific focus on the HABs history during the last ca 2500 yrs was attempted. Age control was performed with five radiocarbon dates obtained from benthic foraminifera. A total number of 71 dinoflagellate cyst types was recorded. The most common types found consistently throughout the sediment sequence were autotrophic Spiniferites spp., Spiniferites hyperacanthus and S. mirabilis, S. ramosus, Operculodinium centrocarpum sensu Wall and Dale 1966, Polysphaeridium zoharyi, and heterotrophic Brigantedinium ssp., cyst of Gymnodinium catenatum and cysts mixture of Protoperidinium. Three local dinoflagellate zones LAX-1 to LAX-3 were established based on the results of the constrained cluster analysis and data ordination; additionally, the middle zone LAX-2 was derived into two subzones, LAX-2a and LAX-2b based on the dynamics of toxic and heterotrophic cysts as well as on the significant changes (probability, P=0.89) in percentages of eutrophic indicators. The total cyst count varied from 106 to 410 dinocysts per slide, with 177 cyst types on average. Dinocyst assemblages are characterized by high values of the dost-depositional degradation index (kt) that varies between 3.6 and 7.6 (averaging 5.4), which is relatively high and is very typical for the areas with selective dinoflagellate cyst preservation that is related to bottom-water oxygen concentrations.

Keywords: reconstruction of palaeoenvironment, harmful algal blooms, anthropogenic influence on coastal zones, South China Sea

Procedia PDF Downloads 90

Authors: Zeynep Tanriverdi̇

Abstract:

Sustainable architecture is a design approach that provides healthy, comfortable, safe, clean space production as well as utilizes minimum resources for efficient and economical use of natural resources and energy. Traditional houses located in rural areas are sustainable structures built at the design and implementation stage in accordance with the climatic environmental data of the region and also effectively using natural energy resources. The fact that these structures are located in an earthquake geography like Türkiye brings their earthquake resistance to the agenda. Since the construction of these structures, which contain the architectural and technological cultural knowledge of the past, is shaped according to the characteristics of the regions where they are located, their resistance to earthquakes also differs. Analyses in rural areas after the earthquake show that there are light-damaged structures that can survive, severely damaged structures, and completely destroyed structures. In this regard, experts can implement repair, consolidation, and reconstruction applications, respectively. While simple repair interventions are carried out in accordance with the original data in traditional houses that have shown great resistance to earthquakes, reinforcement work blended with new technologies can be applied in damaged structures. In reconstruction work, a wide variety of applications can be seen with the possibilities of modern technologies. In rural areas experiencing earthquakes around the world, there are experimental new housing applications that are renewable, environmentally friendly, and sustainable with modern construction techniques in the light of scientific data. With these new residences, it is aimed to create earthquake-resistant, economical, healthy, and pain-relieving therapy spaces for people whose daily lives have been interrupted by disasters. In this study, the preservation of high earthquake-prone rural areas will be discussed through the knowledge transfer of traditional architecture and also permanent housing practices using new sustainable technologies to improve the area. In this way, it will be possible to keep losses to a minimum with sustainable, reliable applications prepared for the worst aspects of the disaster situation and to establish a link between the knowledge of the past and the new technologies of the future.

Keywords: sustainability, conservation, traditional construction systems and materials, new technologies, earthquake resistance

Procedia PDF Downloads 66

Authors: Hossein Aghababaei, Sergio Vitale, Giampaolo Ferraioli

Abstract:

In the context of Synthetic Aperture Radar (SAR) data, polarization is an important source of information for Earth's surface monitoring. SAR Systems are often considered to transmit only one polarization. This constraint leads to either single or dual polarimetric SAR imaging modalities. Single polarimetric systems operate with a fixed single polarization of both transmitted and received electromagnetic (EM) waves, resulting in a single acquisition channel. Dual polarimetric systems, on the other hand, transmit in one fixed polarization and receive in two orthogonal polarizations, resulting in two acquisition channels. Dual polarimetric systems are obviously more informative than single polarimetric systems and are increasingly being used for a variety of remote sensing applications. In dual polarimetric systems, the choice of polarizations for the transmitter and the receiver is open. The choice of circular transmit polarization and coherent dual linear receive polarizations forms a special dual polarimetric system called hybrid polarimetry, which brings the properties of rotational invariance to geometrical orientations of features in the scene and optimizes the design of the radar in terms of reliability, mass, and power constraints. The complete characterization of target scattering, however, requires fully polarimetric data, which can be acquired with systems that transmit two orthogonal polarizations. This adds further complexity to data acquisition and shortens the coverage area or swath of fully polarimetric images compared to the swath of dual or hybrid polarimetric images. The search for solutions to augment dual polarimetric data to full polarimetric data will therefore take advantage of full characterization and exploitation of the backscattered field over a wider coverage with less system complexity. Several methods for reconstructing fully polarimetric images using hybrid polarimetric data can be found in the literature. Although the improvements achieved by the newly investigated and experimented reconstruction techniques are undeniable, the existing methods are, however, mostly based upon model assumptions (especially the assumption of reflectance symmetry), which may limit their reliability and applicability to vegetation and forest scenarios. To overcome the problems of these techniques, this paper proposes a new framework for reconstructing fully polarimetric information from hybrid polarimetric data. The framework uses Deep Learning solutions to augment hybrid polarimetric data without relying on model assumptions. A convolutional neural network (CNN) with a specific architecture and loss function is defined for this augmentation problem by focusing on different scattering properties of the polarimetric data. In particular, the method controls the CNN training process with respect to several characteristic features of polarimetric images defined by the combination of different terms in the cost or loss function. The proposed method is experimentally validated with real data sets and compared with a well-known and standard approach from the literature. From the experiments, the reconstruction performance of the proposed framework is superior to conventional reconstruction methods. The pseudo fully polarimetric data reconstructed by the proposed method also agree well with the actual fully polarimetric images acquired by radar systems, confirming the reliability and efficiency of the proposed method.

Keywords: SAR image, polarimetric SAR image, convolutional neural network, deep learnig, deep neural network

Procedia PDF Downloads 71

Authors: Hossein Aghababaei, Sergio Vitale, Giampaolo ferraioli

Abstract:

In the context of Synthetic Aperture Radar (SAR) data, polarization is an important source of information for Earth's surface monitoring . SAR Systems are often considered to transmit only one polarization. This constraint leads to either single or dual polarimetric SAR imaging modalities. Single polarimetric systems operate with a fixed single polarization of both transmitted and received electromagnetic (EM) waves, resulting in a single acquisition channel. Dual polarimetric systems, on the other hand, transmit in one fixed polarization and receive in two orthogonal polarizations, resulting in two acquisition channels. Dual polarimetric systems are obviously more informative than single polarimetric systems and are increasingly being used for a variety of remote sensing applications. In dual polarimetric systems, the choice of polarizations for the transmitter and the receiver is open. The choice of circular transmit polarization and coherent dual linear receive polarizations forms a special dual polarimetric system called hybrid polarimetry, which brings the properties of rotational invariance to geometrical orientations of features in the scene and optimizes the design of the radar in terms of reliability, mass, and power constraints. The complete characterization of target scattering, however, requires fully polarimetric data, which can be acquired with systems that transmit two orthogonal polarizations. This adds further complexity to data acquisition and shortens the coverage area or swath of fully polarimetric images compared to the swath of dual or hybrid polarimetric images. The search for solutions to augment dual polarimetric data to full polarimetric data will therefore take advantage of full characterization and exploitation of the backscattered field over a wider coverage with less system complexity. Several methods for reconstructing fully polarimetric images using hybrid polarimetric data can be found in the literature. Although the improvements achieved by the newly investigated and experimented reconstruction techniques are undeniable, the existing methods are, however, mostly based upon model assumptions (especially the assumption of reflectance symmetry), which may limit their reliability and applicability to vegetation and forest scenarios. To overcome the problems of these techniques, this paper proposes a new framework for reconstructing fully polarimetric information from hybrid polarimetric data. The framework uses Deep Learning solutions to augment hybrid polarimetric data without relying on model assumptions. A convolutional neural network (CNN) with a specific architecture and loss function is defined for this augmentation problem by focusing on different scattering properties of the polarimetric data. In particular, the method controls the CNN training process with respect to several characteristic features of polarimetric images defined by the combination of different terms in the cost or loss function. The proposed method is experimentally validated with real data sets and compared with a well-known and standard approach from the literature. From the experiments, the reconstruction performance of the proposed framework is superior to conventional reconstruction methods. The pseudo fully polarimetric data reconstructed by the proposed method also agree well with the actual fully polarimetric images acquired by radar systems, confirming the reliability and efficiency of the proposed method.

Keywords: SAR image, deep learning, convolutional neural network, deep neural network, SAR polarimetry

Procedia PDF Downloads 93

Authors: Minghao Liu

Abstract:

This paper uses the theory of communication in the context of mass communication, from the construction of communication symbols, communication flow organization, communication experience perception of the three levels of the old industrial factory landscape transformation research and analysis, summarizes the old industrial factory landscape in the communication process to create strategies and design methods for the old industrial factories carried by the urban culture of how to enter the public's life more widely in the existing environment and be familiar with the significance of the exploration, to provide a new idea for the renewal of the urban stock, and ultimately to achieve the sustainable development of the city.

Keywords: communication, old industrial factor, urban renewal, landscape design

Procedia PDF Downloads 101

Authors: Inga Žukovaitė

Abstract:

This work aims to present postdoc research on remote criminal proceedings in court in order to streamline the proceedings and, at the same time, ensure the effective participation of the parties in criminal proceedings and the court's obligation to administer substantive and procedural justice. This study tests the hypothesis that remote criminal proceedings do not in themselves violate the fundamental principles of criminal procedure; however, their implementation must ensure the right of the parties to effective legal remedies and a fair trial and, only then, must address the issues of procedural economy, speed and flexibility/functionality of the application of technologies. In order to ensure that changes in the regulation of criminal proceedings are in line with fair trial standards, this research will provide answers to the questions of what conditions -first of all, legal and only then organisational- are required for remote criminal proceedings to ensure respect for the parties and enable their effective participation in public proceedings, to create conditions for quality legal defence and its accessibility, to give a correct impression to the party that they are heard and that the court is impartial and fair. It also seeks to present the results of empirical research in the courts of Lithuania that was made by using the interview method. The research will serve as a basis for developing a theoretical model for remote criminal proceedings in the EU to ensure a balance between the intention to have innovative, cost-effective, and flexible criminal proceedings and the positive obligation of the State to ensure the rights of participants in proceedings to just and fair criminal proceedings. Moreover, developments in criminal proceedings also keep changing the image of the court itself; therefore, in the paper will create preconditions for future research on the impact of remote criminal proceedings on the trust in courts. The study aims at laying down the fundamentals for theoretical models of a remote hearing in criminal proceedings and at making recommendations for the safeguarding of human rights, in particular the rights of the accused, in such proceedings. The following criteria are relevant for the remote form of criminal proceedings: the purpose of judicial instance, the legal position of participants in proceedings, their vulnerability, and the nature of required legal protection. The content of the study consists of: 1. Identification of the factual and legal prerequisites for a decision to organise the entire criminal proceedings by remote means or to carry out one or several procedural actions by remote means 2. After analysing the legal regulation and practice concerning the application of the elements of remote criminal proceedings, distinguish the main legal safeguards for protection of the rights of the accused to ensure: (a) the right of effective participation in a court hearing; (b) the right of confidential consultation with the defence counsel; (c) the right of participation in the examination of evidence, in particular material evidence, as well as the right to question witnesses; and (d) the right to a public trial.

Keywords: remote criminal proceedings, fair trial, right to defence, technology progress

Procedia PDF Downloads 73

Authors: Alireza Rahpeyma

Abstract:

One of the most significant challenges faced by cities is inefficient and deteriorated urban fabric. Deteriorated areas bring about numerous issues, including economic, social, physical, and infrastructural problems, sewage management, environmental concerns, and security issues. One of the crucial necessities of modern urban life is the revitalization and renovation of these urban fabrics. Another important aspect is preserving a cohesive cultural and social identity during the urban renewal process. Urban renovation and upgrading are not a one-time occurrence but rather an ongoing process that ultimately needs to become ingrained in the system. This process is not swift; it involves sensitive and intricate stages that require well-designed plans within short, medium, and long-term timeframes. To revitalize dilapidated areas, a comprehensive understanding of the urban region's sustainability is essential. The goal of this study is to provide a suitable model for the upgrading and renewal of areas 2, 4, 5, 7, 8, and 10 in Tehran, preserving the heritage within these urban fabrics (urban assets), including the valuable morphologies of these areas, to prevent financial wastage. This study was conducted descriptively and analytically using the SWOT technique and GIS software. The research results regarding urban upgrading and renewal bring about the following points: 1) Urban upgrading and renewal ensure satisfactory efficiency, justice, environmental quality, safety and security, hygiene, sewage management, and the comfort and aesthetics of urban spaces. 2) Employing the above processes in old city centers re-establishes the historical and cultural connection between people and their past, a past that is evolving and improving and will not repeat itself. 3) Despite the expansion and growth of Iranian cities, cultural and national identity can be preserved by adhering to healthy urban revitalization principles. 4) Proper urban upgrading and renewal prevent social fragmentation, deviant behavioral patterns, psychological and occupational disorders, and financial losses. In general, preserving the identity and urban life in renovation and upgrading must be comprehensive, encompassing various dimensions, including physical, social, economic, political, and administrative aspects.

Keywords: dilapidated areas, renovation, sustainable reconstruction, Tehran-Iran

Procedia PDF Downloads 56

Authors: Antonio Oliveira, Rute F. Meneses, Nuno Trigueiros-Cunha

Abstract:

Tinnitus is commonly defined as an aberrant perception of sound without external stimulus. It is a chronic condition, with consequences on the QOL. The coping strategies used were not always effective and coping was identified as a predictor of QOL in individuals with tinnitus, which reinforces the idea that in health the use of effective coping styles should be promoted. This work intend to verify relations between coping strategies assessed by BriefCope in subjects with tinnitus and variables such as gender, age and severity of tinnitus measured by THI and the Visual Analogue Scale and also hearing and hyperacusis. The results indicate that there are any statistically significant relationships between the variables assessed in relation to the results of BriefCope except in the Visual Analogue Scale. These results, indicating no relationship between almost all variables, reinforce the need for further study of coping strategies use by these patients.

Keywords: Brief Cope, coping strategies, quality of live, THI, Tinnitus

Procedia PDF Downloads 520

Authors: Noman Shahzad, Amyn Pardhan, Hasnain Zafar

Abstract:

Introduction: Though short term survival advantage of damage control laparotomy in management of critically ill trauma patients is established, there is little known about the long-term quality of life of these patients. Facial closure rate after damage control laparotomy is reported to be 20-70 percent. Abdominal wall reconstruction in those who failed to achieve facial closure is challenging and can potentially affect quality of life of these patients. Methodology: We conducted retrospective matched cohort study. Adult patients who underwent damage control laparotomy from Jan 2007 till Jun 2013 were identified through medical record. Patients who had concomitant disabling brain injury or limb injuries requiring amputation were excluded. Age, gender and presentation time matched non exposure group of patients who underwent laparotomy for trauma but no damage control were identified for each damage control laparotomy patient. Quality of life assessment was done via telephonic interview at least one year after the operation, using Urdu version of EuroQol Group Quality of Life (QOL) questionnaire EQ5D after permission. Wilcoxon signed rank test was used to compare QOL scores and McNemar test was used to compare individual parameters of QOL questionnaire. Study was approved by institutional ethical review committee. Results: Out of 32 patients who underwent damage control laparotomy during study period, 20 fulfilled the selection criteria for which 20 matched controls were selected. Median age of patients (IQ Range) was 33 (26-40) years. Facial closure rate in damage control laparotomy group was 40% (8/20). One third of those who did not achieve facial closure (4/12) underwent abdominal wall reconstruction. Self-reported QOL score of damage control laparotomy patients was significantly worse than non-damage control group (p = 0.032). There was no statistically significant difference in two groups regarding individual QOL measures. Significantly, more patients in damage control group were requiring use of abdominal binder, and more patients in damage control group had to either change their job or had limitations in continuing previous job. Our study was not adequately powered to detect factors responsible for worse QOL in damage control group. Conclusion: Quality of life of damage control patients is worse than their age and gender matched patients who underwent trauma laparotomy but not damage control. Adequately powered studies need to be conducted to explore factors responsible for this finding for potential improvement.

Keywords: damage control laparotomy, laparostomy, quality of life

Procedia PDF Downloads 280

Authors: Faizan Sultan

Abstract:

Non-government organizations (NGOs) in Pakistan have the potential to deliver services such as health, education, and rural development through targeting the most vulnerable communities of society. Having this significant importance, NGOs are facing numerous challenges in service delivery. So, there is a need to identify the challenges NGOs face in community development, particularly post-conflict development. The current study has analyzed the social and political constraints in development projects in the post-conflict region of the Swat district of Khyber Pakhtunkhwa. The objectives of this study are “What are the social and political constraints faced by the nongovernmental organizations in the implementation of development aid Projects in post-conflict development of Swat and to examine the challenges in coordination mechanism between government departments, NGOs, and community in reconstruction activities”. This research is based upon both the quantitative and qualitative data that is being gathered from the NGO representatives, government officials, and community members who were involved in post-conflict development interventions in the Swat region. A purposive sampling technique was used to select respondents from the community members/activists (25 in number) and government and NGO officials (10 in number). Based on analysis against our objectives, NGOs have faced numerous constraints such as Insecurity, Negative Perceptions about NGOs, restrictions on women's mobility, government policies and regulations, lack of coordination and networking, trust deficit, and political interference while implementing their project interventions. These findings concluded that constraints have affected project implementation to a greater extent, including women's participation, involvement of marginalized populations, and equal distribution of resources. In the Swat region, NGOs cannot openly discuss sensitive projects such as human rights, gender-based projects, or women empowerment as these issues are very sensitive to the local community due to their cultural values. The community may not allow their females to go outside their homes as this region is a male-dominated society. Similarly, lack of communication and poor networking for the arrangements of the project meetings were also the major constraints.

Keywords: national disaster management authority, millennium development goals, provincial disaster management authority, provincial reconstruction, rehabilitation and settlement authority

Procedia PDF Downloads 59

Authors: Theresa Chinyere ONU

Abstract:

This paper examines the effect of social terrorism on education in Nigeria. The article looked into some prevailing conditions of international political unrest and insecurity. The fear and risk of these conditions to national security and the struggle for power establishment which has further intensified and taken the shape of terrorism has imposed devastating effects on the growth and prosperity of Nigeria; as traffic patterns get disturbed, hospitals and schools get dysfunctional. This has also affected the educational standard in Nigeria as parents are no longer comfortable in sending their children to schools in some states for the fear of terrorist attacks. The study emphasized the integrated the effort of the government management institutions.

Keywords: education, social terrorism, women, Nigeria

Procedia PDF Downloads 584

Authors: Snehal G. Teli, R. J. Shelke

Abstract:

CNN and MultiUNet models are the framework for the proposed method for enhancing and reconstructing underwater images. Multiscale merging of features and regeneration are both performed by the MultiUNet. CNN collects relevant features. Extensive tests on benchmark datasets show that the proposed strategy performs better than the latest methods. As a result of this work, underwater images can be represented and interpreted in a number of underwater applications with greater clarity. This strategy will advance underwater exploration and marine research by enhancing real-time underwater image processing systems, underwater robotic vision, and underwater surveillance.

Keywords: convolutional neural network, image enhancement, machine learning, multiunet, underwater images

Procedia PDF Downloads 78

Authors: Essam Al Daoud

Abstract:

This study solves a phylogeny problem by using modified wild dog pack optimization. The least squares error is considered as a cost function that needs to be minimized. Therefore, in each iteration, new distance matrices based on the constructed trees are calculated and used to select the alpha dog. To test the suggested algorithm, ten homologous genes are selected and collected from National Center for Biotechnology Information (NCBI) databanks (i.e., 16S, 18S, 28S, Cox 1, ITS1, ITS2, ETS, ATPB, Hsp90, and STN). The data are divided into three categories: 50 taxa, 100 taxa and 500 taxa. The empirical results show that the proposed algorithm is more reliable and accurate than other implemented methods.

Keywords: least square, neighbor joining, phylogenetic tree, wild dog pack

Procedia PDF Downloads 320

Authors: Majed A. Alsalem

Abstract:

Universal Design for Learning (UDL) is common theme in education across the US and an influential model and framework that enables students in general and particularly students who are deaf and hard of hearing (DHH) to access the general education curriculum. UDL helps teachers determine how information will be presented to students and how to keep students engaged. Moreover, UDL helps students to express their understanding and knowledge to others. UDL relies on technology to promote students' interaction with content and their communication of knowledge. This study included 120 DHH students who received daily instruction based on UDL principles. This study presents the results of the study and discusses its implications for the integration of UDL in day-to-day practice as well as in the country's education policy. UDL is a Western concept that began and grew in the US, and it has just begun to transfer to other countries such as Saudi Arabia. It will be very important to researchers, practitioners, and educators to see how UDL is being implemented in a new place with a different culture. UDL is a framework that is built to provide multiple means of engagement, representation, and action and expression that should be part of curricula and lessons for all students. The purpose of this study is to investigate the variables associated with the implementation of UDL in Saudi Arabian schools and identify the barriers that could prevent the implementation of UDL. Therefore, this study used a mixed methods design that use both quantitative and qualitative methods. More insights will be gained by including both quantitative and qualitative rather than using a single method. By having methods that different concepts and approaches, the databases will be enriched. This study uses levels of collecting date through two stages in order to insure that the data comes from multiple ways to mitigate validity threats and establishing trustworthiness in the findings. The rationale and significance of this study is that it will be the first known research that targets UDL in Saudi Arabia. Furthermore, it will deal with UDL in depth to set the path for further studies in the Middle East. From a perspective of content, this study considers teachers’ implementation knowledge, skills, and concerns of implementation. This study deals with effective instructional designs that have not been presented in any conferences, workshops, teacher preparation and professional development programs in Saudi Arabia. Specifically, Saudi Arabian schools are challenged to design inclusive schools and practices as well as to support all students’ academic skills development. The total participants in stage one were 336 teachers of DHH students. The results of the intervention indicated significant differences among teachers before and after taking the training sessions associated with their understanding and level of concern. Teachers have indicated interest in knowing more about UDL and adopting it into their practices; they reported that UDL has benefits that will enhance their performance for supporting student learning.

Keywords: deaf and hard of hearing, professional development, Saudi Arabia, universal design for learning

Procedia PDF Downloads 432

Authors: Mansi Chinchanikar

Abstract:

Homoeopathy is the second most widely used medical system in the world, yet the homoeopaths of India and around the world are sick of reading or hearing about how homoeopathy is only a placebo effect and cannot cure or even manage any disease. However, individuals making such unfounded claims should explain to the group how a homoeopathic placebo, particularly one for a neurodevelopmental disease like Attention Deficit Hyperactivity Disorder (ADHD), can be effective in children, with studies to back it up their skeptics. This literary review work exhibits how homoeopathy with a multimodal approach may show a considerable proportion of ADHD patients in India and throughout the world successfully manageable and treatable according to growing study evidence, ruling out the hazardous conventional medicines. Indeed, homeopathy can help cure ADHD symptoms either on its own or in combination with other types of integrative systems.

Keywords: ADHD, adult ADHD, homoeopathy, integrative approach

Procedia PDF Downloads 82

Authors: Kelum N. Manamendra-Arachchi, Kalangi Rodrigo

Abstract:

Bones, teeth, and shells have been acknowledged over the last two centuries as evidence of chronology, Palaeo-environment, and human activity. Faunal traces are valid evidence of past situations because they have properties that have not changed over long periods of time. Sri Lanka has been known as an Island, which has a diverse variation of prehistoric occupation among ecological zones. Defining the Paleoecology of the past societies has been an archaeological thought developed in the 1960s. It is mainly concerned with the reconstruction from available geological and biological evidence of past biota, populations, communities, landscapes, environments, and ecosystems. Sri Lanka has dealt with this subject and considerable research has been already undertaken. The fossil and material record of Sri Lanka’s Wet Zone tropical forests continues from c. 38,000–34,000 ybp. This early and persistent human fossil, technical, and cultural florescence, as well as a collection of well-preserved tropical-forest rock shelters with associated ' on-site ' Palaeoenvironmental records, makes Sri Lanka a central and unusual case study to determine the extent and strength of early human tropical forest encounters. Excavations carried out in prehistoric caves in the low country wet zone has shown that in the last 50,000 years, the temperature in the lowland rainforests has not exceeded 5 degrees. Based on Semnopithecus Priam (Gray Langur) remains unearned from wet zone prehistoric caves, it has been argued that periods of momentous climate changes during the LGM and Terminal Pleistocene/Early Holocene boundary, with a recognizable preference for semi-open ‘Intermediate’ rainforest or edges. Continuous Genus Acavus and Oligospira occupation along with uninterrupted horizontal pervasive of Canarium sp. (‘kekuna’ nut) have proven that temperatures in the lowland rain forests have not changed by at least 5 oC over the last 50,000 years. Site Catchment or Territorial analysis cannot be no longer defensible, due to time-distance based factors as well as optimal foraging theory failed as a consequences of prehistoric people were aware of the decrease in cost-benefit ratio and located sites, and generally played out a settlement strategy that minimized the ratio of energy expanded to energy produced.

Keywords: palaeo-environment, prehistory, palaeo-ecology, zooarchaeology

Procedia PDF Downloads 123

Authors: Jiankang Wang, Hongyang Yu

Abstract:

This paper proposes a portable online 3D modeling method. The method first utilizes a depth camera to collect data and compresses the depth data using a frame-by-frame lossless data compression method. The color image is encoded using the H.264 encoding format. After the cloud obtains the color image and depth image, a 3D modeling method based on bundlefusion is used to complete the 3D modeling. The results of this study indicate that this method has the characteristics of portability, online, and high efficiency and has a wide range of application prospects.

Keywords: 3D reconstruction, bundlefusion, lossless compression, depth image

Procedia PDF Downloads 82

Authors: Lei Hou, Xinli Du, Nikolaos Boulgouris

Abstract:

A cochlear implant, referred to as a CI, is a small electronic device that can provide direct electrical stimulation to the auditory nerve. During cochlear implant surgery, atraumatic electrode array insertion is considered to be a crucial step. However, during implantation, the mechanical behaviour of an electrode array inside the cochlea is not known. The behaviour of an electrode array inside of the cochlea is hardly identified by regular methods. In this study, a CI electrode array capacitive sensor system is proposed. It is able to automatically determine the array state as a result of the capacitance variations. Instead of applying sensors to the electrode array, the capacitance information from the electrodes will be gathered and analysed. Results reveal that this sensing method is capable of recognising different states when fed into a pre-shaped model.

Keywords: cochlear implant, electrode, hearing preservation, insertion force, capacitive sensing

Procedia PDF Downloads 239

Authors: Chia-Yu Lin, Yu-Hsiang Shen, Cing-Ciao Ke, Chia-Hao Chang, Fan-Pin Tseng, Yu-Ching Ni, Sheng-Pin Tseng

Abstract:

Tomosynthesis is an imaging system that generates a 3D image by scanning in a limited angular range. It could provide more depth information than traditional 2D X-ray single projection. Radiation dose in tomosynthesis is less than computed tomography (CT). Because of limited angular range scanning, there are many properties depending on scanning direction. Therefore, non-coplanar imaging technique was developed to improve image quality in traditional tomosynthesis. The purpose of this study was to establish the non-coplanar imaging technique of tomosynthesis system and evaluate this technique by the reconstructed image. INER prototype tomosynthesis system contains an X-ray tube, a flat panel detector, and a motion machine. This system could move X-ray tube in multiple directions during the acquisition. In this study, we investigated three different imaging techniques that were 2D X-ray single projection, traditional tomosynthesis, and non-coplanar tomosynthesis. An anthropopathic chest phantom was used to evaluate the image quality. It contained three different size lesions (3 mm, 5 mm and, 8 mm diameter). The traditional tomosynthesis acquired 61 projections over a 30 degrees angular range in one scanning direction. The non-coplanar tomosynthesis acquired 62 projections over 30 degrees angular range in two scanning directions. A 3D image was reconstructed by iterative image reconstruction algorithm (ML-EM). Our qualitative method was to evaluate artifacts in tomosynthesis reconstructed image. The quantitative method was used to calculate a peak-to-valley ratio (PVR) that means the intensity ratio of the lesion to the background. We used PVRs to evaluate the contrast of lesions. The qualitative results showed that in the reconstructed image of non-coplanar scanning, anatomic structures of chest and lesions could be identified clearly and no significant artifacts of scanning direction dependent could be discovered. In 2D X-ray single projection, anatomic structures overlapped and lesions could not be discovered. In traditional tomosynthesis image, anatomic structures and lesions could be identified clearly, but there were many artifacts of scanning direction dependent. The quantitative results of PVRs show that there were no significant differences between non-coplanar tomosynthesis and traditional tomosynthesis. The PVRs of the non-coplanar technique were slightly higher than traditional technique in 5 mm and 8 mm lesions. In non-coplanar tomosynthesis, artifacts of scanning direction dependent could be reduced and PVRs of lesions were not decreased. The reconstructed image was more isotropic uniformity in non-coplanar tomosynthesis than in traditional tomosynthesis. In the future, scan strategy and scan time will be the challenges of non-coplanar imaging technique.

Keywords: image reconstruction, non-coplanar imaging technique, tomosynthesis, X-ray imaging

Procedia PDF Downloads 370

Authors: Jean-Eric Pelet, Erhard Lick, Basma Taieb

Abstract:

Purpose This research bridges the gap between sensory marketing and the use of the Internet of Things (IoT) in luxury hotels. We investigated how stimulating guests’ senses through IoT devices influenced their emotions, affective experiences, eudaimonism (well-being), and, ultimately, guest behavior. We examined potential moderating effects of gender. Design/methodology/approach We adopted a mixed method approach, combining qualitative research (semi-structured interviews) to explore hotel managers’ perspectives on the potential use of IoT in luxury hotels and quantitative research (surveying hotel guests; n=357). Findings The results showed that while the senses of smell, hearing, and sight had an impact on guests’ emotions, the senses of touch, hearing, and sight impacted guests’ affective experiences. The senses of smell and taste influenced guests’ eudaimonism. The sense of smell had a greater effect on eudaimonism and behavioral intentions among women compared to men. Originality IoT can be applied in creating customized multi-sensory hotel experiences. For example, hotels may offer unique and diverse ambiences in their rooms and suites to improve guest experiences. Research limitations/implications This study concentrated on luxury hotels located in Europe. Further research may explore the generalizability of the findings (e.g., in other cultures, comparison between high-end and low-end hotels). Practical implications Context awareness and hyper-personalization, through intensive and continuous data collection (hyper-connectivity) and real time processing, are key trends in the service industry. Therefore, big data plays a crucial role in the collection of information since it allows hoteliers to retrieve, analyze, and visualize data to provide personalized services in real time. Together with their guests, hotels may co-create customized sensory experiences. For instance, if the hotel knows about the guest’s music preferences based on social media as well as their age and gender, etc. and considers the temperature and size (standard, suite, etc.) of the guest room, this may determine the playlist of the concierge-tablet made available in the guest room. Furthermore, one may record the guest’s voice to use it for voice command purposes once the guest arrives at the hotel. Based on our finding that the sense of smell has a greater impact on eudaimonism and behavioral intentions among women than men, hotels may deploy subtler scents with lower intensities, or even different scents, for female guests in comparison to male guests.

Keywords: affective experience, emotional value, eudaimonism, hospitality industry, Internet of Things, sensory marketing

Procedia PDF Downloads 57

Authors: Dariusz Jacek Jakóbczak

Abstract:

Mathematics and computer science are interested in methods of 2D curve interpolation and extrapolation using the set of key points (knots). A proposed method of Hurwitz- Radon Matrices (MHR) is such a method. This novel method is based on the family of Hurwitz-Radon (HR) matrices which possess columns composed of orthogonal vectors. Two-dimensional curve is interpolated via different functions as probability distribution functions: polynomial, sinus, cosine, tangent, cotangent, logarithm, exponent, arcsin, arccos, arctan, arcctg or power function, also inverse functions. It is shown how to build the orthogonal matrix operator and how to use it in a process of curve reconstruction.

Keywords: 2D data interpolation, hurwitz-radon matrices, MHR method, probabilistic modeling, curve extrapolation

Procedia PDF Downloads 525

Authors: Aljoša Šorgo

Abstract:

Recent scholarship on Phrygian has begun to more closely examine the long-held belief that Greek and Phrygian are two very closely related languages. It is now clear that Graeco-Phrygian can be firmly postulated as a subclade of the Indo-European languages. The present paper will focus on the reconstruction of the phonological and phonetic segments of Proto-Graeco-Phrygian (= PGPh.) by providing relevant correspondence sets and reconstructing the classes of segments. The PGPh. basic vowel system consisted of ten phonemic oral vowels: */a e o ā ē ī ō ū/. The correspondences of the vowels are clear and leave little open to ambiguity. There were four resonants and two semi-vowels in PGPh.: */r l m n i̯ u̯/, which could appear in both a consonantal and a syllabic function, with the distribution between the two still being phonotactically predictable. Of note is the fact that the segments *m and *n seem to have merged when their phonotactic position would see them used in a syllabic function. Whether the segment resulting from this merger was a nasalized vowel (most likely *[ã]) or a syllabic nasal *[N̥] (underspecified for place of articulation) cannot be determined at this stage. There were three fricatives in PGPh.: */s h ç/. *s and *h are easily identifiable. The existence of *ç, which may seem unexpected, is postulated on the basis of the correspondence Gr. ὄς ~ Phr. yos/ιος. It is of note that Bozzone has previously proposed the existence of *ç ( < PIE *h₁i̯-) in an early stage of Greek even without taking into account Phrygian data. Finally, the system of stops in PGPh. distinguished four places of articulation (labial, dental, velar, and labiovelar) and three phonation types. The question of which three phonation types were actually present in PGPh. is one of great importance for the ongoing debate on the realization of the three series in PIE. Since the matter is still very much in dispute, we ought to, at this stage, endeavour to reconstruct the PGPh. system without recourse to the other IE languages. The three series of correspondences are: 1. Gr. T (= tenuis) ~ Phr. T; 2. Gr. D (= media) ~ Phr. T; 3. Gr. TA (= tenuis aspirata) ~ Phr. M. The first series must clearly be reconstructed as composed of voiceless stops. The second and third series are more problematic. With a bottom-up approach, neither the second nor the third series of correspondences are compatible with simple modal voicing, and the reflexes differ greatly in voice onset time. Rather, the defining feature distinguishing the two series was [±spread glottis], with ancillary vibration of the vocal cords. In PGPh. the second series was undergoing further spreading of the glottis. As the two languages split, this process would continue, but be affected by dissimilar changes in VOT, which was ultimately phonemicized in both languages as the defining feature distinguishing between their series of stops.

Keywords: bottom-up reconstruction, Proto-Graeco-Phrygian, spread glottis, syllabic resonant

Procedia PDF Downloads 50

Authors: Ora Geiger

Abstract:

Reading is a cognitive process of deciphering visual signs to produce meaning. During the reading process, written information of symbols and signs is received in the person’s eye and processed in the brain. This definition is relevant to both the reading of letters and the reading of musical notation. But while the letters of the alphabet are signs determined arbitrarily, notes are recorded systematically on a staff, with the location of each note on the staff indicating its relative pitch. In this paper, the researcher specifies the characteristics of alphabet reading in comparison to musical notation reading, and discusses the question whether a person diagnosed with dyslexia will necessarily have difficulty in reading musical notes. Dyslexia is a learning disorder that makes it difficult to acquire alphabet-reading skills due to difficulties expressed in the identification of letters, spelling, and other language deciphering skills. In order to read, one must be able to connect a symbol with a sound and to join the sounds into words. A person who has dyslexia finds it difficult to translate a graphic symbol into the sound that it represents. When teaching reading to children diagnosed with dyslexia, the multi-sensory approach, supporting the activation and involvement of most of the senses in the learning process, has been found to be particularly effective. According to this approach, when most senses participate in the reading learning process, it becomes more effective. During years of experience, the researcher, who is a music specialist, has been following the music reading learning process of elementary school age students, some of them diagnosed with Dyslexia, while studying to play soprano (descant) recorder. She argues that learning music reading while studying to play a musical instrument is a multi-sensory experience by its nature. The senses involved are: sight, hearing, touch, and the kinesthetic sense (motion), which provides the brain with information on the relative positions of the body. In this way, the learner experiences simultaneously visual, auditory, tactile, and kinesthetic impressions. The researcher concludes that there should be no contra-indication for teaching standard music reading to children with dyslexia if an appropriate process is offered. This conclusion is based on two main characteristics of music reading: (1) musical notation system is a systematic, logical, relative set of symbols written on a staff; and (2) music reading learning connected with playing a musical instrument is by its nature a multi-sensory activity since it combines sight, hearing, touch, and movement. This paper describes music reading teaching procedures and provides unique teaching methods that have been found to be effective for students who were diagnosed with Dyslexia. It provides theoretical explanations in addition to guidelines for music education practices.

Keywords: alphabet reading, dyslexia, multisensory teaching method, music reading, recorder playing

Procedia PDF Downloads 366

Authors: Siavash Mirahmadizoghi, Steven Bell, David Simpson

Abstract:

In this work a new independent component analysis (ICA) based method for noise reduction in evoked potentials is evaluated on for auditory late responses (ALR) captured with a 63-channel electroencephalogram (EEG) from 10 normal-hearing subjects. The performance of the new method is compared with a single channel alternative in terms of signal to noise ratio (SNR), the number of channels with an SNR above an empirically derived statistical critical value and an estimate of the effect of attention on the major components in the ALR waveform. The results show that the multichannel signal processing method can significantly enhance the quality of the ALR signal and also detect the effect of the attention on the ALR better than the single channel alternative.

Keywords: auditory late response (ALR), attention, EEG, independent component analysis (ICA), multichannel signal processing

Procedia PDF Downloads 505

Authors: Hyohun Kim, Dongwha Shin, Yeonseok Kim, Ji-Su Ahn, Kensuke Nakamura, Dongeun Choi, Byung-Woo Hong

Abstract:

Artefacts are commonly encountered in the imaging process of clinical computed tomography (CT) where the artefact refers to any systematic discrepancy between the reconstructed observation and the true attenuation coefficient of the object. It is known that CT images are inherently more prone to artefacts due to its image formation process where a large number of independent detectors are involved, and they are assumed to yield consistent measurements. There are a number of different artefact types including noise, beam hardening, scatter, pseudo-enhancement, motion, helical, ring, and metal artefacts, which cause serious difficulties in reading images. Thus, it is desired to remove nuisance factors from the degraded image leaving the fundamental intrinsic information that can provide better interpretation of the anatomical and pathological characteristics. However, it is considered as a difficult task due to the high dimensionality and variability of data to be recovered, which naturally motivates the use of machine learning techniques. We propose an image restoration algorithm based on the deep neural network framework where the denoising auto-encoders are stacked building multiple layers. The denoising auto-encoder is a variant of a classical auto-encoder that takes an input data and maps it to a hidden representation through a deterministic mapping using a non-linear activation function. The latent representation is then mapped back into a reconstruction the size of which is the same as the size of the input data. The reconstruction error can be measured by the traditional squared error assuming the residual follows a normal distribution. In addition to the designed loss function, an effective regularization scheme using residual-driven dropout determined based on the gradient at each layer. The optimal weights are computed by the classical stochastic gradient descent algorithm combined with the back-propagation algorithm. In our algorithm, we initially decompose an input image into its intrinsic representation and the nuisance factors including artefacts based on the classical Total Variation problem that can be efficiently optimized by the convex optimization algorithm such as primal-dual method. The intrinsic forms of the input images are provided to the deep denosing auto-encoders with their original forms in the training phase. In the testing phase, a given image is first decomposed into the intrinsic form and then provided to the trained network to obtain its reconstruction. We apply our algorithm to the restoration of the corrupted CT images by the artefacts. It is shown that our algorithm improves the readability and enhances the anatomical and pathological properties of the object. The quantitative evaluation is performed in terms of the PSNR, and the qualitative evaluation provides significant improvement in reading images despite degrading artefacts. The experimental results indicate the potential of our algorithm as a prior solution to the image interpretation tasks in a variety of medical imaging applications. This work was supported by the MISP(Ministry of Science and ICT), Korea, under the National Program for Excellence in SW (20170001000011001) supervised by the IITP(Institute for Information and Communications Technology Promotion).

Keywords: auto-encoder neural network, CT image artefact, deep learning, intrinsic image representation, noise reduction, total variation

Procedia PDF Downloads 190