Search results for: successive images method

Authors: C. Castro-Egler, A. Durán-Escalante, A. García-González

Abstract:

This paper presents a method to generate a finite element model of the human auditory inner ear system. The geometric model has been realized using 2D images from a virtual model of temporal bones. A point cloud has been gotten manually from those images to construct a whole mesh with hexahedral elements. The main difference with the predecessor models is the spiral shape of the cochlea with its three scales completely defined: scala tympani, scala media and scala vestibuli; which are separate by basilar membrane and Reissner membrane. To validate this model, numerical simulations have been realised with two models: an isolated inner ear and a whole model of human auditory system. Ideal conditions of displacement are applied over the oval window in the isolated Inner Ear model. The whole model is made up of the outer auditory channel, the tympani, the ossicular chain, and the inner ear. The boundary condition for the whole model is 1Pa over the auditory channel entrance. The numerical simulations by FEM have been done using a harmonic analysis with a frequency range between 100-10.000 Hz with an interval of 100Hz. The following results have been carried out: basilar membrane displacement; the scala media pressure according to the cochlea length and the transfer function of the middle ear normalized with the pressure in the tympanic membrane. The basilar membrane displacements and the pressure in the scala media make it possible to validate the response in frequency of the basilar membrane.

Keywords: finite elements method, human auditory system model, numerical analysis, 3D modelling cochlea

Procedia PDF Downloads 361

Authors: Firas Gerges, Frank Y. Shih

Abstract:

Malignant melanoma, known simply as melanoma, is a type of skin cancer that appears as a mole on the skin. It is critical to detect this cancer at an early stage because it can spread across the body and may lead to the patient's death. When detected early, melanoma is curable. In this paper, we propose a deep learning model (convolutional neural networks) in order to automatically classify skin lesion images as malignant or benign. Images underwent certain pre-processing steps to diminish the effect of the normal skin region on the model. The result of the proposed model showed a significant improvement over previous work, achieving an accuracy of 97%.

Keywords: deep learning, skin cancer, image processing, melanoma

Procedia PDF Downloads 146

Authors: Uma Maheshwari V., Rajanikanth Aluvalu, Kumar Gautam

Abstract:

The COVID-19 disease is a highly contagious viral infection with major worldwide health implications. The global economy suffers as a result of COVID. The spread of this pandemic disease can be slowed if positive patients are found early. COVID-19 disease prediction is beneficial for identifying patients' health problems that are at risk for COVID. Deep learning and machine learning algorithms for COVID prediction using X-rays have the potential to be extremely useful in solving the scarcity of doctors and clinicians in remote places. In this paper, a convolutional neural network (CNN) with deep layers is presented for recognizing COVID-19 patients using real-world datasets. We gathered around 6000 X-ray scan images from various sources and split them into two categories: normal and COVID-impacted. Our model examines chest X-ray images to recognize such patients. Because X-rays are commonly available and affordable, our findings show that X-ray analysis is effective in COVID diagnosis. The predictions performed well, with an average accuracy of 99% on training photographs and 88% on X-ray test images.

Keywords: deep CNN, COVID–19 analysis, feature extraction, feature map, accuracy

Procedia PDF Downloads 77

Authors: M. G. Smitha, M. V. Murugendrappa

Abstract:

Perovskite manganite La_0.7Ca_0.3MnO₃ was synthesized by Sol-gel method. Polymerization of pyrrole was carried by in-situ polymerization method. The composite of pyrrole (Py)/La_0.7Ca_0.3MnO₃ composite in the presence of oxidizing agent ammonium per sulphate to synthesize polypyrrole (PPy)/La_0.7Ca_0.3MnO₃ (LCM) composite was carried out by the same in-situ polymerization method. The PPy/LCM composites were synthesized with varying compositions like 10, 20, 30, 40, and 50 wt.% of LCM in Py. The surface morphologies of these composites were analyzed by using scanning electron microscope (SEM). The images show that LCM particles are embedded in PPy chain. The impedance measurement of PPy/LCM at different temperature ranges from 30 to 180 °C was studied using impedance analyzer. The study shows that impedance is frequency and temperature dependent and it is found to decrease with increase in frequency and temperature.

Keywords: polypyrrole, sol gel, impedance, composites

Procedia PDF Downloads 371

Authors: Massimo Miccoli, Luca Marangoni, Alberto Aniello Scaringi, Alessandro Marceddu, Alessandro Amicone

Abstract:

The susceptibility of deep neural networks (DNNs) to adversarial manipulations is a recognized challenge within the computer vision domain. Adversarial examples, crafted by adding subtle yet malicious alterations to benign images, exploit this vulnerability. Various defense strategies have been proposed to safeguard DNNs against such attacks, stemming from diverse research hypotheses. Building upon prior work, our approach involves the utilization of autoencoder models. Autoencoders, a type of neural network, are trained to learn representations of training data and reconstruct inputs from these representations, typically minimizing reconstruction errors like mean squared error (MSE). Our autoencoder was trained on a dataset of benign examples; learning features specific to them. Consequently, when presented with significantly perturbed adversarial examples, the autoencoder exhibited high reconstruction errors. The architecture of the autoencoder was tailored to the dimensions of the images under evaluation. We considered various image sizes, constructing models differently for 256x256 and 512x512 images. Moreover, the choice of the computer vision model is crucial, as most adversarial attacks are designed with specific AI structures in mind. To mitigate this, we proposed a method to replace image-specific dimensions with a structure independent of both dimensions and neural network models, thereby enhancing robustness. Our multi-modal autoencoder reconstructs the spectral representation of images across the red-green-blue (RGB) color channels. To validate our approach, we conducted experiments using diverse datasets and subjected them to adversarial attacks using models such as ResNet50 and ViT_L_16 from the torch vision library. The autoencoder extracted features used in a classification model, resulting in an MSE (RGB) of 0.014, a classification accuracy of 97.33%, and a precision of 99%.

Keywords: adversarial attacks, malicious images detector, binary classifier, multimodal transformer autoencoder

Procedia PDF Downloads 111

Authors: Boukhatem Mohammed Belkaid, Lahdir Mourad

Abstract:

In this paper, we propose a new encryption system for security issues medical images. The hybrid encryption scheme is based on AES and RSA algorithms to validate the three security services are authentication, integrity, and confidentiality. Privacy is ensured by AES, authenticity is ensured by the RSA algorithm. Integrity is assured by the basic function of the correlation between adjacent pixels. Our system generates a unique password every new session of encryption, that will be used to encrypt each frame of the medical image basis to strengthen and ensure his safety. Several metrics have been used for various tests of our analysis. For the integrity test, we noticed the efficiencies of our system and how the imprint cryptographic changes at reception if a change affects the image in the transmission channel.

Keywords: AES, RSA, integrity, confidentiality, authentication, medical images, encryption, decryption, key, correlation

Procedia PDF Downloads 538

Authors: Alireza Shayegan, Morteza Amirabadi

Abstract:

The y tool was developed to quantitatively compare dose distributions, either measured or calculated. Before computing ɣ, the dose and distance scales of the two distributions, referred to as evaluated and reference, are re-normalized by dose and distance criteria, respectively. The re-normalization allows the dose distribution comparison to be conducted simultaneously along dose and distance axes. Several two-dimensional images were acquired using a Scanning Liquid Ionization Chamber EPID and Extended Dose Range (EDR2) films for regular and irregular radiation fields. The raw images were then converted into two-dimensional dose maps. Transitional and rotational manipulations were performed for images using Matlab software. As evaluated dose distribution maps, they were then compared with the corresponding original dose maps as the reference dose maps.

Keywords: energetic electron, gamma function, penumbra, Matlab software

Procedia PDF Downloads 298

Authors: Sandhya Rao

Abstract:

Kevin Lynch’s Theory on Imeageability, let on explore a city in terms of five elements, Nodes, Paths, Edges, landmarks and Districts. What happens when we try to record the same data in an Indian context? What happens when we apply the same theory of Imageability to a complex shifting urban pattern of the Indian cities and how can we as Urban Designers demonstrate our role in the image building ordeal of these cities? The organizational patterns formed through mental images, of an Indian city is often diverse and intangible. It is also multi layered and temporary in terms of the spirit of the place. The pattern of images formed is loaded with associative meaning and intrinsically linked with the history and socio-cultural dominance of the place. The embedded memory of a place in one’s mind often plays an even more important role while formulating these images. Thus while deriving an image of a city one is often confused or finds the result chaotic. The images formed due to its complexity are further difficult to represent using a single medium. Under such a scenario it’s difficult to derive an output of an image constructed as well as make design interventions to enhance the legibility of a place. However, there can be a combination of tools and methods that allows one to record the key elements of a place through time, space and one’s user interface with the place. There has to be a clear understanding of the participant groups of a place and their time and period of engagement with the place as well. How we can translate the result obtained into a design intervention at the end, is the main of the research. Could a multi-faceted cognitive mapping be an answer to this or could it be a very transient mapping method which can change over time, place and person. How does the context influence the process of image building in one’s mind? These are the key questions that this research will aim to answer.

Keywords: imageability, organizational patterns, legibility, cognitive mapping

Procedia PDF Downloads 312

Authors: Nour-Elhouda Lecheheb

Abstract:

Destination image is one of the main factors that influence potential visitors' decision choice. This study aims to explore the pre-visit perception of prior British tourists and compare them to the actual projected images of the Algerian tourism suppliers. Semi-structured interviews are conducted with both prior British tourists to Algeria and the Algerian tourism suppliers in 2019. The findings of this study suggest how the Algerian tourism suppliers might benefit from understanding the perceived image of prior tourists to match tourists' expectations and better plan their projected images.

Keywords: Algeria, destination choice, destination image, perceived image, projected image

Procedia PDF Downloads 165

Authors: M. O. Bayram, H. B. Gencdal, N. O. Fercan, B. Basbug

Abstract:

Liquid limit, which is used as a measure of soil strength, can be detected by Casagrande and fall-cone testing methods. The two methods majorly diverge from each other in terms of operator dependency. The Casagrande method that is applied according to ASTM D4318-17 standards may give misleading results, especially if the calibration process is not performed well. To reveal the effect of calibration for drop height and amount of soil paste placement in the Casagrande cup, a series of tests were carried out by multipoint method as it is specified in the ASTM standards. The tests include the combination of 6 mm, 8 mm, 10 mm, and 12 mm drop heights and under-filled, half-filled, and full-filled Casagrande cups by kaolinite samples. It was observed that during successive tests, the drop height of the cup deteriorated; hence the device was recalibrated before and after each test to provide the accuracy of the results. Besides, the tests by under-filled and full-filled samples for higher drop heights revealed lower liquid limit values than the lower drop heights revealed. For the half-filled samples, it was clearly seen that the liquid limit values didn’t change at all as the drop height increased, and this explains the function of standard specifications.

Keywords: calibration, casagrande cup method, drop height, kaolinite, liquid limit, placing form

Procedia PDF Downloads 158

Authors: Nadia Ben Youssef, Aicha Bouzid

Abstract:

Gradient estimation is one of the most fundamental tasks in the field of image processing in general, and more particularly for color images since that the research in color image gradient remains limited. The widely used gradient method is Di Zenzo’s gradient operator, which is based on the measure of squared local contrast of color images. The proposed gradient mechanism, presented in this paper, is based on the principle of the Di Zenzo’s approach using quaternion representation. This edge detector is compared to a marginal approach based on multiscale product of wavelet transform and another vector approach based on quaternion convolution and vector gradient approach. The experimental results indicate that the proposed color gradient operator outperforms marginal approach, however, it is less efficient then the second vector approach.

Keywords: gradient, edge detection, color image, quaternion

Procedia PDF Downloads 232

Authors: Cristina Maria Ribeiro Martins Pereira Caridade, Ana Rita Ferreira Morais

Abstract:

The use of computational tools has become essential in the context of interactive learning, especially in engineering education. In the medical industry, teaching medical image processing techniques is a crucial part of training biomedical engineers, as it has integrated applications with healthcare facilities and hospitals. The aim of this article is to present a teaching-learning simulation tool developed in MATLAB using a graphical user interface for medical image fusion that explores different image fusion methodologies and processes in combination with image pre-processing techniques. The application uses different algorithms and medical fusion techniques in real time, allowing you to view original images and fusion images, compare processed and original images, adjust parameters, and save images. The tool proposed in an innovative teaching and learning environment consists of a dynamic and motivating teaching simulation for biomedical engineering students to acquire knowledge about medical image fusion techniques and necessary skills for the training of biomedical engineers. In conclusion, the developed simulation tool provides real-time visualization of the original and fusion images and the possibility to test, evaluate and progress the student’s knowledge about the fusion of medical images. It also facilitates the exploration of medical imaging applications, specifically image fusion, which is critical in the medical industry. Teachers and students can make adjustments and/or create new functions, making the simulation environment adaptable to new techniques and methodologies.

Keywords: image fusion, image processing, teaching-learning simulation tool, biomedical engineering education

Procedia PDF Downloads 129

Authors: Javaid Marium, Ahmad Iftikhar

Abstract:

This study was aimed to identify factors that affect negative mood and body image dissatisfaction in women. positive and negative affect, self esteem, body image satisfaction and figure rating scale was administered to 97 female undergraduate students. This served as a base line data for correlation analysis in the first instance. One week later participants who volunteered to appear in the second phase of the study (N=47) were shown thin- ideal images as an intervention and soon after they completed positive and negative affect schedule and body image states scale again as a post test. Results indicated body mass as a strong negative predictor of body image dis/satisfaction, self esteem was a moderate predictor and mood was not a significant predictor. The participants whose actual body shape was markedly discrepant with the ideally desired body shape had significantly low level of body image satisfaction (p < .001) than those with low discrepancy. Similar results were found for self esteem (p < .004). Both self esteem and body mass predicted body satisfaction about equally and significantly. However, on viewing thin-ideal images, the participants of different body weight showed no change in their body image satisfaction than before. Only the overweight participants were significantly affected on negative mood as a short term reaction after viewing the thin ideal images. Comparing the three groups based on their body mass, one-way ANOVA revealed significant difference on negative mood as well as body image satisfaction. This reveals body mass as a potent and stable factor that consistently and strongly affected body satisfaction not the transient portrayal of thin ideal images.

Keywords: body image satisfaction, thin-ideal images, media, mood affects, self esteem

Procedia PDF Downloads 281

Authors: D. P. Gangwar

Abstract:

Identification of human through facial images has got great importance in forensic science. The video recordings, CCTV footage, passports, driver licenses and other related documents are invariably sent to the laboratory for comparison of the questioned photographs as well as video recordings with suspected photographs/recordings to prove the identity of a person. More than 300 questioned and 300 control photographs received in actual crime cases, received from various investigation agencies, have been compared by me so far using various familiar analysis and comparison techniques such as Holistic comparison, Morphological analysis, Photo-anthropometry and superimposition. On the basis of findings obtained during the examination huge photo exhibits, a realistic and comprehensive technique has been proposed which could be very useful for forensic.

Keywords: CCTV Images, facial features, photo-anthropometry, superimposition

Procedia PDF Downloads 527

Authors: Saad Algharib, Jay Lee

Abstract:

Urbanization is a complex phenomenon that occurs during the city’s development from one form to another. In other words, it is the process when the activities in the land use/land cover change from rural to urban. Since the oil exploration, Kuwait City has been growing rapidly due to its urbanization and population growth by both natural growth and inward immigration. The main objective of this study is to detect changes in urban land use/land cover and to examine the changing spatial patterns of urban growth in and around Kuwait City between 1989 and 2001. In addition, this study also evaluates the spatial patterns of the changes detected and how they can be related to the spatial configuration of the city. Recently, the use of remote sensing and geographic information systems became very useful and important tools in urban studies because of the integration of them can allow and provide the analysts and planners to detect, monitor and analyze the urban growth in a region effectively. Moreover, both planners and users can predict the trends of the growth in urban areas in the future with remotely sensed and GIS data because they can be effectively updated with required precision levels. In order to identify the new urban areas between 1989 and 2001, the study uses satellite images of the study area and remote sensing technology for classifying these images. Unsupervised classification method was applied to classify images to land use and land cover data layers. After finishing the unsupervised classification method, GIS overlay function was applied to the classified images for detecting the locations and patterns of the new urban areas that developed during the study period. GIS was also utilized to evaluate the distribution of the spatial patterns. For example, Moran’s index was applied for all data inputs to examine the urban growth distribution. Furthermore, this study assesses if the spatial patterns and process of these changes take place in a random fashion or with certain identifiable trends. During the study period, the result of this study indicates that the urban growth has occurred and expanded 10% from 32.4% in 1989 to 42.4% in 2001. Also, the results revealed that the largest increase of the urban area occurred between the major highways after the forth ring road from the center of Kuwait City. Moreover, the spatial distribution of urban growth occurred in cluster manners.

Keywords: geographic information systems, remote sensing, urbanization, urban growth

Procedia PDF Downloads 170

Authors: Saad M. Darwish, Mohamed A. El-Iskandarani, Guitar M. Shawkat

Abstract:

Nowadays, the amount of available multimedia data is continuously on the rise. The need to find a required image for an ordinary user is a challenging task. Content based image retrieval (CBIR) computes relevance based on the visual similarity of low-level image features such as color, textures, etc. However, there is a gap between low-level visual features and semantic meanings required by applications. The typical method of bridging the semantic gap is through the automatic image annotation (AIA) that extracts semantic features using machine learning techniques. In this paper, a multi-label image annotation system guided by Firefly and Bayesian method is proposed. Firstly, images are segmented using the maximum variance intra cluster and Firefly algorithm, which is a swarm-based approach with high convergence speed, less computation rate and search for the optimal multiple threshold. Feature extraction techniques based on color features and region properties are applied to obtain the representative features. After that, the images are annotated using translation model based on the Net Bayes system, which is efficient for multi-label learning with high precision and less complexity. Experiments are performed using Corel Database. The results show that the proposed system is better than traditional ones for automatic image annotation and retrieval.

Keywords: feature extraction, feature selection, image annotation, classification

Procedia PDF Downloads 584

Authors: H. K. Bonsu-Owu

Abstract:

Media portrayal of women in traditional stereotypical roles such as mothers, or seductress has been the norm for years. However, the changing socioeconomic and political environment and advancement of women in today’s society have given rise to questions on the appropriate portrayal of women in the media today. The purpose of the study is to analyze the portrayal of women in Ghanaian newspapers and find women’s perception on the issue. The study uses a three step approach in gathering data for analysis. Using the stratified sampling method, it analyzes front page images of women from 210 issues of the selected newspapers. Further, it administers questionnaires to 100 female students to find out how they relate to the images of women in the selected newspapers. Finally, editors of the newspapers are interviewed to find their rational for portraying women as seen on their front pages. The findings suggest that the newspapers portray women for varied reasons such as promoting sales and influencing the public agenda. Further, the female students claim that in spite of women’s vast contribution to the growth of society, the media continue to marginalize them. They add that such portrayals promote and reinforce social construct, however, refuse to see themselves through the male gaze concept. The study concludes that the stereotyped portrayal of women is likely to continue if the government, regulatory bodies, the media and society do not make a conscious effort to address this problem.

Keywords: women, newspaper, portrayal, social construct

Procedia PDF Downloads 132

Authors: Yavuz Unal, Kemal Polat, H. Erdinc Kocer

Abstract:

In this study, a decision support system comprising three stages has been proposed to detect the disc abnormalities of the lumbar region. In the first stage named the feature extraction, T2-weighted sagittal and axial Magnetic Resonance Images (MRI) were taken from 55 people and then 27 appearance and shape features were acquired from both sagittal and transverse images. In the second stage named the feature weighting process, k-means clustering based feature weighting (KMCBFW) proposed by Gunes et al. Finally, in the third stage named the classification process, the classifier algorithms including multi-layer perceptron (MLP- neural network), support vector machine (SVM), Naïve Bayes, and decision tree have been used to classify whether the subject has lumbar disc or not. In order to test the performance of the proposed method, the classification accuracy (%), sensitivity, specificity, precision, recall, f-measure, kappa value, and computation times have been used. The best hybrid model is the combination of k-means clustering based feature weighting and decision tree in the detecting of lumbar disc disease based on both sagittal and axial MR images.

Keywords: lumbar disc abnormality, lumbar MRI, lumbar spine, hybrid models, hybrid features, k-means clustering based feature weighting

Procedia PDF Downloads 517

Authors: Fayzan Ahmed, Saad Bin Saeed, Abdul Qadir Jangda

Abstract:

Gafchromic™ sheet are extensively used for the QA of intensity modulated radiation therapy and other in-vivo dosimetry. Intra-sheet Non-uniformity of scanner as well as film causes undesirable fluctuations which are reflected in dosimetry The aim of this study is to define a systematic and robust method to investigate the intra-sheet uniformity of the unexposed Gafchromic Sheets and the region of interest (ROI) of the scanner. Sheets of lot No#: A05151201 were scanned before and after the expiry period with the EPSON™ XL10000 scanner in the transmission mode, landscape orientation and 72 dpi resolution. ROI of (8’x 10’ inches) equal to the sheet dimension in the center of the scanner is used to acquire images with full transmission, block transmission and with sheets in place. 500 virtual grids, created in MATALB® are imported as a macros in ImageJ (1.49m Wayne Rasband) to analyze the images. In order to remove the edge effects, the outer 86 grids are excluded from the analysis. The standard deviation of the block transmission and full transmission are 0.38% and 0.66% confirming a higher uniformity of the scanner. Expired and non-expired sheets have standard deviations of 2.18% and 1.29%, show that uniformity decreases after expiry. The results are promising and indicates a good potential of this method to be used as a uniformity check for scanner and unexposed Gafchromic sheets.

Keywords: IMRT, film dosimetry, virtual grids, uniformity

Procedia PDF Downloads 490

Authors: R. Daira

Abstract:

The electronic speckle interferometry technique used to measure the deformations of scatterers process is based on the subtraction of interference patterns. A speckle image is first recorded before deformation of the object in the RAM of a computer, after a second deflection. The square of the difference between two images showing correlation fringes observable in real time directly on monitor. The interpretation these fringes to determine the deformation. In this paper, we present experimental results of deformation out of the plane of two samples in aluminum, electronic boards and stainless steel.

Keywords: optical method, holography, interferometry, deformation

Procedia PDF Downloads 404

Authors: Abdelkrim Belhaoua, Jean-Pierre Radoux, Mariana Kuras, Vincent Récamier, Martial Piotto, Karim Elbayed, François Proust, Izzie Namer

Abstract:

This paper proposes an augmented reality system dedicated to neurosurgery in order to assist the surgeon during an operation. This work is part of the ExtempoRMN project (Funded by Bpifrance) which aims at analyzing during a surgical operation the metabolic content of tumoral brain biopsy specimens by HRMAS NMR. Patients affected with a brain tumor (gliomas) frequently need to undergo an operation in order to remove the tumoral mass. During the operation, the neurosurgeon removes biopsy specimens using image-guided surgery. The biopsy specimens removed are then sent for HRMAS NMR analysis in order to obtain a better diagnosis and prognosis. Image-guided refers to the use of MRI images and a computer to precisely locate and target a lesion (abnormal tissue) within the brain. This is performed using preoperative MRI images and the BrainLab neuro-navigation system. With the patient MRI images loaded on the Brainlab Cranial neuro-navigation system in the operating theater, surgeons can better identify their approach before making an incision. The Brainlab neuro-navigation tool tracks in real time the position of the instruments and displays their position on the patient MRI data. The results of the biopsy analysis by 1H HRMAS NMR are then sent back to the operating theater and superimposed on the 3D localization system directly on the MRI images. The method we have developed to communicate between the HRMAS NMR analysis software and Brainlab makes use of a combination of C++, VTK and the Insight Toolkit using OpenIGTLink protocol.

Keywords: neuro-navigation, augmented reality, biopsy, BrainLab, HR-MAS NMR

Procedia PDF Downloads 362

Authors: Javier Sedano-Cibrián, Julio Manuel de Luis-Ruiz, Rubén Pérez-Álvarez, Raúl Pereda-García, Beatriz Malagón-Picón

Abstract:

Structural inspection activities are necessary to ensure the correct functioning of infrastructures. Unmanned Aerial Vehicle (UAV) techniques have become more popular than traditional techniques. Specifically, UAV Photogrammetry allows time and cost savings. The development of this technology has permitted the use of low-cost thermal sensors in UAVs. The representation of 3D thermal models with this type of equipment is in continuous evolution. The direct processing of thermal images usually leads to errors and inaccurate results. A methodology is proposed for the generation of 3D thermal models using dual sensors, which involves the application of visible Red-Blue-Green (RGB) and thermal images in parallel. Hence, the RGB images are used as the basis for the generation of the model geometry, and the thermal images are the source of the surface temperature information that is projected onto the model. Mining/industrial facilities representations that are obtained can be used for inspection activities.

Keywords: aerial thermography, data processing, drone, low-cost, point cloud

Procedia PDF Downloads 142

Authors: Kamal KC, Zhendong Yin, Dasen Li, Zhilu Wu

Abstract:

Intra-plant disease classification based on leaf images is a challenging computer vision task due to similarities in texture, color, and shape of leaves with a slight variation of leaf spot; and external environmental changes such as lighting and background noises. Deep convolutional neural network (DCNN) has proven to be an effective tool for binary classification. In this paper, two methods for binary classification of diseased plant leaves using DCNN are presented; model created from scratch and transfer learning. Our main contribution is a thorough evaluation of 4 networks created from scratch and transfer learning of 5 pre-trained models. Training and testing of these models were performed on a plant leaf images dataset belonging to 16 distinct classes, containing a total of 22,265 images from 8 different plants, consisting of a pair of healthy and diseased leaves. We introduce a deep CNN model, Optimized MobileNet. This model with depthwise separable CNN as a building block attained an average test accuracy of 99.77%. We also present a fine-tuning method by introducing the concept of a convolutional block, which is a collection of different deep neural layers. Fine-tuned models proved to be efficient in terms of accuracy and computational cost. Fine-tuned MobileNet achieved an average test accuracy of 99.89% on 8 pairs of [healthy, diseased] leaf ImageSet.

Keywords: deep convolution neural network, depthwise separable convolution, fine-grained classification, MobileNet, plant disease, transfer learning

Procedia PDF Downloads 185

Authors: Mohammad Tarek Al Muallim, Ozhan Duzenli, Ceyhun Ilguy

Abstract:

Harbor protection systems are so important. The need for automatic protection systems has increased over the last years. Diver detection active sonar has great significance. It used to detect underwater threats such as divers and autonomous underwater vehicle. To automatically detect such threats the sonar image is processed by algorithms. These algorithms used to detect, track and classify of underwater objects. In this work, divers tracking and classification algorithm is improved be proposing a robust wake detection method. To detect objects the sonar images is normalized then segmented based on fixed threshold. Next, the centroids of the segments are found and clustered based on distance metric. Then to track the objects linear Kalman filter is applied. To reduce effect of noise and creation of false tracks, the Kalman tracker is fine tuned. The tuning is done based on our active sonar specifications. After the tracks are initialed and updated they are subjected to a filtering stage to eliminate the noisy and unstable tracks. Also to eliminate object with a speed out of the diver speed range such as buoys and fast boats. Afterwards the result tracks are subjected to a classification stage to deiced the type of the object been tracked. Here the classification stage is to deice wither if the tracked object is an open circuit diver or a close circuit diver. At the classification stage, a small area around the object is extracted and a novel wake detection method is applied. The morphological features of the object with his wake is extracted. We used support vector machine to find the best classifier. The sonar training images and the test images are collected by ARMELSAN Defense Technologies Company using the portable diver detection sonar ARAS-2023. After applying the algorithm to the test sonar data, we get fine and stable tracks of the divers. The total classification accuracy achieved with the diver type is 97%.

Keywords: harbor protection, diver detection, active sonar, wake detection, diver classification

Procedia PDF Downloads 237

Authors: Jianhong Xiang, Hao Xiang, Linyu Wang

Abstract:

Compressed sensing (CS) has a wide range of applications in sparse signal reconstruction. Aiming at the problems of low recovery accuracy and long reconstruction time of existing reconstruction algorithms in medical imaging, this paper proposes a corrected smoothing L0 algorithm based on compressed sensing (CSL0). First, an approximate hyperbolic tangent function (AHTF) that is more similar to the L0 norm is proposed to approximate the L0 norm. Secondly, in view of the "sawtooth phenomenon" in the steepest descent method and the problem of sensitivity to the initial value selection in the modified Newton method, the use of the steepest descent method and the modified Newton method are jointly optimized to improve the reconstruction accuracy. Finally, the CSL0 algorithm is simulated on various images. The results show that the algorithm proposed in this paper improves the reconstruction accuracy of the test image by 0-0. 98dB.

Keywords: smoothed L0, compressed sensing, image processing, sparse reconstruction

Procedia PDF Downloads 113

Authors: El Asnaoui Khalid, Chawki Youness, Aksasse Brahim, Ouanan Mohammed

Abstract:

In these recent decades, the important and fast growth in the development and demand of multimedia products is contributing to an insufficient in the bandwidth of device and network storage memory. Consequently, the theory of data compression becomes more significant for reducing the data redundancy in order to save more transfer and storage of data. In this context, this paper addresses the problem of the lossless and the near-lossless compression of images. This proposed method is based on Block SVD Power Method that overcomes the disadvantages of Matlab's SVD function. The experimental results show that the proposed algorithm has a better compression performance compared with the existing compression algorithms that use the Matlab's SVD function. In addition, the proposed approach is simple and can provide different degrees of error resilience, which gives, in a short execution time, a better image compression.

Keywords: image compression, SVD, block SVD power method, lossless compression, near lossless

Procedia PDF Downloads 384

Authors: Simon L. Madsen, Mads Dyrmann, Morten S. Laursen, Rasmus N. Jørgensen

Abstract:

Recent advances in the Unmanned Aerial System (UAS) safety and perception systems enable safe low altitude autonomous terrain following flights recently demonstrated by the consumer DJI Mavic PRO and Phamtom 4 Pro drones. This paper presents the first prototype system utilizing this functionality in form of semi-automated UAS based collection of crop/weed images where the embedded perception system ensures a significantly safer and faster gathering of weed images with sub-millimeter resolution. The system is to be used when the weeds are at cotyledon stage and prior to the harvest recognizing the grass weed species, which cannot be discriminated at the cotyledon stage.

Keywords: weed mapping, UAV, DJI SDK, automation, cotyledon plants

Procedia PDF Downloads 307

Authors: Lae-Jeong Park, Kyusoo Chung, Jungho Moon

Abstract:

This paper introduces an effective method of segmenting Korean text (place names in Korean) from a Korean road sign image. A Korean advanced directional road sign is composed of several types of visual information such as arrows, place names in Korean and English, and route numbers. Automatic classification of the visual information and extraction of Korean place names from the road sign images make it possible to avoid a lot of manual inputs to a database system for management of road signs nationwide. We propose a series of problem-specific heuristics that correctly segments Korean place names, which is the most crucial information, from the other information by leaving out non-text information effectively. The experimental results with a dataset of 368 road sign images show 96% of the detection rate per Korean place name and 84% per road sign image.

Keywords: segmentation, road signs, characters, classification

Procedia PDF Downloads 442

Authors: R. Sumalatha, M. V. Subramanyam

Abstract:

In medical image processing the quality of the image is degraded in the presence of noise. Especially in ultra sound imaging and Magnetic resonance imaging the data was corrupted by signal dependent noise known as salt and pepper noise. Removal of noise from the medical images is a critical issue for researchers. In this paper, a new type of technique Adaptive Spatial Mask Filter (ASMF) has been proposed. The proposed filter is used to increase the quality of MRI and ultra sound images. Experimental results show that the proposed filter outperforms the implementation of mean, median, adaptive median filters in terms of MSE and PSNR.

Keywords: salt and pepper noise, ASMF, PSNR, MSE

Procedia PDF Downloads 434

Authors: Heewon Jeong, Seongpyo Kim, Joon Ha Kim

Abstract:

Over the last few tens of years, the coastal regions of Korea have experienced red tide algal blooms, which are harmful and toxic to both humans and marine organisms due to their potential threat. It was accelerated owing to eutrophication by human activities, certain oceanic processes, and climate change. Previous studies have tried to monitoring and predicting the algae concentration of the ocean with the bio-optical algorithms applied to color images of the satellite. However, the accurate estimation of algal blooms remains problems to challenges because of the complexity of coastal waters. Therefore, this study suggests a new method to identify the concentration of red tide algal bloom from images of geostationary ocean color imager (GOCI) which are representing the water environment of the sea in Korea. The method employed GOCI images, which took the water leaving radiances centered at 443nm, 490nm and 660nm respectively, as well as observed weather data (i.e., humidity, temperature and atmospheric pressure) for the database to apply optical characteristics of algae and train deep learning algorithm. Convolution neural network (CNN) was used to extract the significant features from the images. And then artificial neural network (ANN) was used to estimate the concentration of algae from the extracted features. For training of the deep learning model, backpropagation learning strategy is developed. The established methods were tested and compared with the performances of GOCI data processing system (GDPS), which is based on standard image processing algorithms and optical algorithms. The model had better performance to estimate algae concentration than the GDPS which is impossible to estimate greater than 5mg/m³. Thus, deep learning model trained successfully to assess algae concentration in spite of the complexity of water environment. Furthermore, the results of this system and methodology can be used to improve the performances of remote sensing. Acknowledgement: This work was supported by the 'Climate Technology Development and Application' research project (#K07731) through a grant provided by GIST in 2017.

Keywords: deep learning, algae concentration, remote sensing, satellite

Procedia PDF Downloads 182