Search results for: scale invariant feature

Authors: Jianghao Hu, Hongyu Wang

Abstract:

The effective and efficient human pose estimation algorithm is an important task for real-time human pose estimation on mobile devices. This paper proposes a light-weight human key points detection algorithm, Light-Weight Network for Real-Time Pose Estimation (LWPE). LWPE uses light-weight backbone network and depthwise separable convolutions to reduce parameters and lower latency. LWPE uses the feature pyramid network (FPN) to fuse the high-resolution, semantically weak features with the low-resolution, semantically strong features. In the meantime, with multi-scale prediction, the predicted result by the low-resolution feature map is stacked to the adjacent higher-resolution feature map to intermediately monitor the network and continuously refine the results. At the last step, the key point coordinates predicted in the highest-resolution are used as the final output of the network. For the key-points that are difficult to predict, LWPE adopts the online hard key points mining strategy to focus on the key points that hard predicting. The proposed algorithm achieves excellent performance in the single-person dataset selected in the AI (artificial intelligence) challenge dataset. The algorithm maintains high-precision performance even though the model only contains 3.9M parameters, and it can run at 225 frames per second (FPS) on the generic graphics processing unit (GPU).

Keywords: depthwise separable convolutions, feature pyramid network, human pose estimation, light-weight backbone

Procedia PDF Downloads 122

Authors: Huang Bai-Cheng

Abstract:

When deploying the Oriented Fast and Rotated Binary Robust Independent Elementary Feature (BRIEF) (ORB) extraction algorithm on field-programmable gate arrays (FPGA), the access of global storage for 31×31 pixel patches of the features has become the bottleneck of the system efficiency. Therefore, a feature zone strategy has been proposed. Zones are searched as features are detected. Pixels around the feature zones are extracted from global memory and distributed into patches corresponding to feature coordinates. The proposed FPGA structure is targeted on a Xilinx FPGA development board of Zynq UltraScale+ series, and multiple datasets are tested. Compared with the streaming pixel patch extraction method, the proposed architecture obtains at least two times acceleration consuming extra 3.82% Flip-Flops (FFs) and 7.78% Look-Up Tables (LUTs). Compared with the non-streaming one, the proposed architecture saves 22.3% LUT and 1.82% FF, causing a latency of only 0.2ms and a drop in frame rate for 1. Compared with the related works, the proposed strategy and hardware architecture have the superiority of keeping a balance between FPGA resources and performance.

Keywords: feature extraction, real-time, ORB, FPGA implementation

Procedia PDF Downloads 90

Authors: A: Annis Fathima, V. Vaidehi, S. Ajitha

Abstract:

Face recognition system finds many applications in surveillance and human computer interaction systems. As the applications using face recognition systems are of much importance and demand more accuracy, more robustness in the face recognition system is expected with less computation time. In this paper, a hybrid approach for face recognition combining Gabor Wavelet and Linear Discriminant Analysis (HGWLDA) is proposed. The normalized input grayscale image is approximated and reduced in dimension to lower the processing overhead for Gabor filters. This image is convolved with bank of Gabor filters with varying scales and orientations. LDA, a subspace analysis techniques are used to reduce the intra-class space and maximize the inter-class space. The techniques used are 2-dimensional Linear Discriminant Analysis (2D-LDA), 2-dimensional bidirectional LDA ((2D)2LDA), Weighted 2-dimensional bidirectional Linear Discriminant Analysis (Wt (2D)2 LDA). LDA reduces the feature dimension by extracting the features with greater variance. k-Nearest Neighbour (k-NN) classifier is used to classify and recognize the test image by comparing its feature with each of the training set features. The HGWLDA approach is robust against illumination conditions as the Gabor features are illumination invariant. This approach also aims at a better recognition rate using less number of features for varying expressions. The performance of the proposed HGWLDA approaches is evaluated using AT&T database, MIT-India face database and faces94 database. It is found that the proposed HGWLDA approach provides better results than the existing Gabor approach.

Keywords: face recognition, Gabor wavelet, LDA, k-NN classifier

Procedia PDF Downloads 447

Authors: Tudor Barbu

Abstract:

An automatic moment-based texture segmentation approach is proposed in this paper. First, we describe the related work in this computer vision domain. Our texture feature extraction, the first part of the texture recognition process, produces a set of moment-based feature vectors. For each image pixel, a texture feature vector is computed as a sequence of area moments. Second, an automatic pixel classification approach is proposed. The feature vectors are clustered using some unsupervised classification algorithm, the optimal number of clusters being determined using a measure based on validation indexes. From the resulted pixel classes one determines easily the desired texture regions of the image.

Keywords: image segmentation, moment-based, texture analysis, automatic classification, validation indexes

Procedia PDF Downloads 386

Authors: Sidra Naeem, Ayesha Naeem, Sahar Rahim, Nadia Nawaz Qadri

Abstract:

Citrus greening is a bacterial disease that causes considerable damage to citrus fruits worldwide. Efficient method for this disease detection must be carried out to minimize the production loss. This paper presents a pattern recognition system that comprises three stages for the detection of citrus greening from Orange leaves: segmentation, feature extraction and classification. Image segmentation is accomplished by adaptive thresholding. The feature extraction stage comprises of three visual descriptors i.e. shape, color and texture. From shape feature we have used asymmetry index, from color feature we have used histogram of Cb component from YCbCr domain and from texture feature we have used local binary pattern. Classification was done using support vector machines and k nearest neighbors. The best performances of the system is Accuracy = 88.02% and AUROC = 90.1% was achieved by automatic segmented images. Our experiments validate that: (1). Segmentation is an imperative preprocessing step for computer assisted diagnosis of citrus greening, and (2). The combination of shape, color and texture features form a complementary set towards the identification of citrus greening disease.

Keywords: citrus greening, pattern recognition, feature extraction, classification

Procedia PDF Downloads 144

Authors: Ugochukwu Ejike Akpudo, Jank-Wook Hur

Abstract:

The increasing demand for improved productivity, maintainability, and reliability has prompted rapidly increasing research studies on the emerging condition-based maintenance concept- Prognostics and health management (PHM). Varieties of fuel pumps serve critical functions in several hydraulic systems; hence, their failure can have daunting effects on productivity, safety, etc. The need for condition monitoring and assessment of these pumps cannot be overemphasized, and this has led to the uproar in research studies on standard feature extraction techniques for optimized condition assessment of fuel pumps. By extracting time-based, frequency-based and the more robust time-frequency based features from these vibrational signals, a more comprehensive feature assessment (and selection) can be achieved for a more accurate and reliable condition assessment of these pumps. With the aid of emerging deep classification and regression algorithms like the locally linear embedding (LLE), we propose a method for comprehensive condition assessment of electromagnetic fuel pumps (EMFPs). Results show that the LLE as a comprehensive feature extraction technique yields better feature fusion/dimensionality reduction results for condition assessment of EMFPs against the use of single features. Also, unlike other feature fusion techniques, its capabilities as a fault classification technique were explored, and the results show an acceptable accuracy level using standard performance metrics for evaluation.

Keywords: electromagnetic fuel pumps, comprehensive feature extraction, condition assessment, locally linear embedding, feature fusion

Procedia PDF Downloads 90

Authors: Rohit Gupta, Ravinder Agarwal

Abstract:

Surface electromyographic (sEMG) signal has the potential to identify the human activities and intention. This potential is further exploited to control the artificial limbs using the sEMG signal from residual limbs of amputees. The paper deals with the development of multichannel cost efficient sEMG signal interface for research application, along with evaluation of proposed class dependent statistical approach of the feature selection method. The sEMG signal acquisition interface was developed using ADS1298 of Texas Instruments, which is a front-end interface integrated circuit for ECG application. Further, the sEMG signal is recorded from two lower limb muscles for three locomotions namely: Plane Walk (PW), Stair Ascending (SA), Stair Descending (SD). A class dependent statistical approach is proposed for feature selection and also its performance is compared with 12 preexisting feature vectors. To make the study more extensive, performance of five different types of classifiers are compared. The outcome of the current piece of work proves the suitability of the proposed feature selection algorithm for locomotion recognition, as compared to other existing feature vectors. The SVM Classifier is found as the outperformed classifier among compared classifiers with an average recognition accuracy of 97.40%. Feature vector selection emerges as the most dominant factor affecting the classification performance as it holds 51.51% of the total variance in classification accuracy. The results demonstrate the potentials of the developed sEMG signal acquisition interface along with the proposed feature selection algorithm.

Keywords: classifiers, feature selection, locomotion, sEMG

Procedia PDF Downloads 267

Authors: Nikhat Nawaz, S. Srinivasulu, P. Kesava Rao

Abstract:

Feature extraction plays an important role in many remote sensing applications. Automatic extraction of water bodies is of great significance in many remote sensing applications like change detection, image retrieval etc. This paper presents a procedure for automatic extraction of water information from remote sensing images. The algorithm uses the relative location of R-colour component of the chromaticity diagram. This method is then integrated with the effectiveness of the spatial scale transformation of whole method. The whole method is based on water index fitted from spectral library. Experimental results demonstrate the improved accuracy and effectiveness of the integrated method for automatic extraction of water bodies.

Keywords: feature extraction, remote sensing, image retrieval, chromaticity, water index, spectral library, integrated method

Procedia PDF Downloads 343

Authors: Heramb Somthankar, Anujoy Chakraborty

Abstract:

This paper presents a classification study on yield prediction in semiconductor production using machine learning approaches. A complicated semiconductor production process is generally monitored continuously by signals acquired from sensors and measurement sites. A monitoring system contains a variety of signals, all of which contain useful information, irrelevant information, and noise. In the case of each signal being considered a feature, "Feature Selection" is used to find the most relevant signals. The open-source UCI SECOM Dataset provides 1567 such samples, out of which 104 fail in quality assurance. Feature extraction and selection are performed on the dataset, and useful signals were considered for further study. Afterward, common machine learning algorithms were employed to predict whether the signal yields pass or fail. The most relevant algorithm is selected for prediction based on the accuracy and loss of the ML model.

Keywords: deep learning, feature extraction, feature selection, machine learning classification algorithms, semiconductor production monitoring, signal processing, time-series analysis

Procedia PDF Downloads 76

Authors: Ayşegül Bilge, Merve Uğuryol, Şeyda Dülgerler, Mustafa Yıldız

Abstract:

The purpose of this research is to prove the Attitude towards Using Complementary Treatments Scale reliability and validity. The research is a methodological type of research that has been planned to determine the validity and reliability of the Attitude towards Using Complementary Treatments Scale. The scale has been developed by the researchers. In the scale, there are 23 questions including complementary and modern therapies individuals apply when they have health problems 4-item Likert-type evaluation has been carried out in preparing the questionnaire. High score obtained from the scale indicates a positive attitude towards complementary therapies. In the course of validity assessment of the scale, expert opinion has been received, and the content validity of the scale has been determined by using Kendall coefficient correlation test (Wa=0.200, p = 0.460). In the course of the reliability assessment of the scale, total score correlations of 23 materials have been examined, and those under 0.20 correlation limit has been removed from the scale correlation. As a result, the scale was left to be 13 items. In the internal consistency tests of the analyses, Cronbach's alpha value has been found to be 0.79. As a result, of the validity analyses of the Attitude towards Using Complementary Treatments Scale, the content and language validity analyses has been found to be at the expected level. It has been determined to be a highly reliable scale as the result of the reliability analyses. In conclusion, Attitude towards Using Complementary Treatments Scale is a valid and reliable scale.

Keywords: alternative health care, complementary treatment, instrument development, nursing practice

Procedia PDF Downloads 368

Authors: Yuanchao Sang, Zhihao Gong, Longsheng Chen, Long Chen

Abstract:

Clothing fashion represents a human’s aesthetic appreciation towards everyday outfits and appetite for fashion, and it reflects the development of status in society, humanity, and economics. However, modelling fashion by machine is extremely challenging because fashion is too abstract to be efficiently described by machines. Even human beings can hardly reach a consensus about fashion. In this paper, we are dedicated to answering a fundamental fashion-related problem: what image feature best describes clothing fashion? To address this issue, we have designed and evaluated various image features, ranging from traditional low-level hand-crafted features to mid-level style awareness features to various current popular deep neural network-based features, which have shown state-of-the-art performance in various vision tasks. In summary, we tested the following 9 feature representations: color, texture, shape, style, convolutional neural networks (CNNs), CNNs with distance metric learning (CNNs&DML), AutoEncoder, CNNs with multiple layer combination (CNNs&MLC) and CNNs with dynamic feature clustering (CNNs&DFC). Finally, we validated the performance of these features on two publicly available datasets. Quantitative and qualitative experimental results on both intra-domain and inter-domain fashion clothing image retrieval showed that deep learning based feature representations far outweigh traditional hand-crafted feature representation. Additionally, among all deep learning based methods, CNNs with explicit feature clustering performs best, which shows feature clustering is essential for discriminative fashion feature representation.

Keywords: convolutional neural network, feature representation, image processing, machine modelling

Procedia PDF Downloads 112

Authors: Mohamad Amin Amini, Mehdi Poursha

Abstract:

Nonlinear response history analysis (NL-RHA) is a robust analytical tool for estimating the seismic demands of structures responding in the inelastic range. However, because of its conceptual and numerical complications, the nonlinear static procedure (NSP) is being increasingly used as a suitable tool for seismic performance evaluation of structures. The conventional pushover analysis methods presented in various codes (FEMA 356; Eurocode-8; ATC-40), are limited to the first-mode-dominated structures, and cannot take higher modes effect into consideration. Therefore, since more than a decade ago, researchers developed enhanced pushover analysis procedures to take higher modes effect into account. The main objective of this study is to propose an enhanced invariant lateral displacement distribution to take higher modes effect into consideration in performing a displacement-based pushover analysis, whereby a set of laterally applied displacements, rather than forces, is monotonically applied to the structure. For this purpose, the effect of different parameters such as the spectral displacement of ground motion, the modal participation factor, and the effective modal participating mass ratio on the lateral displacement distribution is investigated to find the best distribution. The major simplification of this procedure is that the effect of higher modes is concentrated into a single invariant lateral load distribution. Therefore, only one pushover analysis is sufficient without any need to utilize a modal combination rule for combining the responses. The invariant lateral displacement distribution for pushover analysis is then calculated by combining the modal story displacements using the modal combination rules. The seismic demands resulting from the different procedures are compared to those from the more accurate nonlinear response history analysis (NL-RHA) as a benchmark solution. Two structures of different heights including 10 and 20-story special steel moment resisting frames (MRFs) were selected and evaluated. Twenty ground motion records were used to conduct the NL-RHA. The results show that more accurate responses can be obtained in comparison with the conventional lateral loads when the enhanced modal lateral displacement distributions are used.

Keywords: displacement-based pushover, enhanced lateral load distribution, higher modes effect, nonlinear response history analysis (NL-RHA)

Procedia PDF Downloads 248

Authors: Pengchao Wei, Xiangzhong Fang

Abstract:

Noise is ubiquitous in this world. Denoising is an essential technology, especially in image semantic segmentation, where noises are generally categorized into two main types i.e. feature noise and label noise. The main focus of this paper is aiming at modeling label noise, investigating the behaviors of different types of label noise on image semantic segmentation tasks using K-Nearest-Neighbor and Convolutional Neural Network classifier. The performance without label noise and with is evaluated and illustrated in this paper. In addition to that, the influence of feature noise on the image semantic segmentation task is researched as well and a feature noise reduction method is applied to mitigate its influence in the learning procedure.

Keywords: convolutional neural network, denoising, feature noise, image semantic segmentation, k-nearest-neighbor, label noise

Procedia PDF Downloads 188

Authors: Weiliang Wen, Xinyu Guo, Ying Zhang, Jianjun Du, Boxiang Xiao

Abstract:

Reconstructing and analyzing three-dimensional (3D) models from situ measured data is important for a number of researches and applications in plant science, including plant phenotyping, functional-structural plant modeling (FSPM), plant germplasm resources protection, agricultural technology popularization. It has many scales like cell, tissue, organ, plant and canopy from micro to macroscopic. The techniques currently used for data capture, feature analysis, and 3D reconstruction are quite different of different scales. In this context, morphological data acquisition, 3D analysis and modeling of plants on different scales are introduced systematically. The commonly used data capture equipment for these multiscale is introduced. Then hot issues and difficulties of different scales are described respectively. Some examples are also given, such as Micron-scale phenotyping quantification and 3D microstructure reconstruction of vascular bundles within maize stalks based on micro-CT scanning, 3D reconstruction of leaf surfaces and feature extraction from point cloud acquired by using 3D handheld scanner, plant modeling by combining parameter driven 3D organ templates. Several application examples by using the 3D models and analysis results of plants are also introduced. A 3D maize canopy was constructed, and light distribution was simulated within the canopy, which was used for the designation of ideal plant type. A grape tree model was constructed from 3D digital and point cloud data, which was used for the production of science content of 11th international conference on grapevine breeding and genetics. By using the tissue models of plants, a Google glass was used to look around visually inside the plant to understand the internal structure of plants. With the development of information technology, 3D data acquisition, and data processing techniques will play a greater role in plant science.

Keywords: plant, three dimensional modeling, multi-scale, plant phenotyping, three dimensional data acquisition

Procedia PDF Downloads 253

Authors: A. Shashank Tiwari, S. P. Mahapatra

Abstract:

Scale-up is defined as the process of increasing batch size. Scale-up of a process viewed as a procedure for applying the same process to different output volumes. There is a subtle difference between these two definitions: batch size enlargement does not always translate into a size increase of the processing volume. In mixing applications, scale-up is indeed concerned with increasing the linear dimensions from the laboratory to the plant size. On the other hand, processes exist (e.g., tableting) where the term ‘scale-up’ simply means enlarging the output by increasing the speed. To complete the picture, one should point out special procedures where an increase of the scale is counterproductive and ‘scale-down’ is required to improve the quality of the product. In moving from Research and Development (R&D) to production scale, it is sometimes essential to have an intermediate batch scale. This is achieved at the so-called pilot scale, which is defined as the manufacturing of drug product by a procedure fully representative of and simulating that used for full manufacturing scale. This scale also makes it possible to produce enough products for clinical testing and to manufacture samples for marketing. However, inserting an intermediate step between R&D and production scales does not, in itself, guarantee a smooth transition. A well-defined process may generate a perfect product both in the laboratory and the pilot plant and then fail quality assurance tests in production.

Keywords: scale up, research, size, batch

Procedia PDF Downloads 377

Authors: Krish Jhurani

Abstract:

The research paper presents an in-depth analysis of symmetry properties in linear algebraic systems under the operation of non-canonical scalar multiplication structures, specifically semirings, and near-rings. The objective is to unveil the profound alterations that occur in traditional linear algebraic structures when we replace conventional field multiplication with these non-canonical operations. In the methodology, we first establish the theoretical foundations of non-canonical scalar multiplication, followed by a meticulous investigation into the resulting symmetry properties, focusing on eigenvectors, eigenspaces, and invariant subspaces. The methodology involves a combination of rigorous mathematical proofs and derivations, supplemented by illustrative examples that exhibit these discovered symmetry properties in tangible mathematical scenarios. The core findings uncover unique symmetry attributes. For linear algebraic systems with semiring scalar multiplication, we reveal eigenvectors and eigenvalues. Systems operating under near-ring scalar multiplication disclose unique invariant subspaces. These discoveries drastically broaden the traditional landscape of symmetry properties in linear algebraic systems. With the application of these findings, potential practical implications span across various fields such as physics, coding theory, and cryptography. They could enhance error detection and correction codes, devise more secure cryptographic algorithms, and even influence theoretical physics. This expansion of applicability accentuates the significance of the presented research. The research paper thus contributes to the mathematical community by bringing forth perspectives on linear algebraic systems and their symmetry properties through the lens of non-canonical scalar multiplication, coupled with an exploration of practical applications.

Keywords: eigenspaces, eigenvectors, invariant subspaces, near-rings, non-canonical scalar multiplication, semirings, symmetry properties

Procedia PDF Downloads 78

Authors: Bechir Hbibi, Lamine Mili

Abstract:

Understanding the condition of comatose patients can be difficult, but it is crucial to their optimal treatment. Consequently, numerous scoring systems have been developed around the world to categorize patient states based on physiological assessments. Although validated and widely adopted by medical communities, these scores still present numerous limitations and obstacles. Even with the addition of additional tests and extensions, these scoring systems have not been able to overcome certain limitations, and it appears unlikely that they will be able to do so in the future. On the other hand, physiological tests are not the only way to extract ideas about comatose patients. EEG signal analysis has helped extensively to understand the human brain and human consciousness and has been used by researchers in the classification of different levels of disease. The use of EEG in the ICU has become an urgent matter in several cases and has been recommended by medical organizations. In this field, the EEG is used to investigate epilepsy, dementia, brain injuries, and many other neurological disorders. It has recently also been used to detect pain activity in some regions of the brain, for the detection of stress levels, and to evaluate sleep quality. In our recent findings, our aim was to use multifractal analysis, a very successful method of handling multifractal signals and feature extraction, to establish a state of awareness scale for comatose patients based on their electrical brain activity. The results show that this score could be instantaneous and could overcome many limitations with which the physiological scales stock. On the contrary, multifractal analysis stands out as a highly effective tool for characterizing non-stationary and self-similar signals. It demonstrates strong performance in extracting the properties of fractal and multifractal data, including signals and images. As such, we leverage this method, along with other features derived from EEG signal recordings from comatose patients, to develop a scale. This scale aims to accurately depict the vigilance state of patients in intensive care units and to address many of the limitations inherent in physiological scales such as the Glasgow Coma Scale (GCS) and the FOUR score. The results of applying version V0 of this approach to 30 patients with known GCS showed that the EEG-based score similarly describes the states of vigilance but distinguishes between the states of 8 sedated patients where the GCS could not be applied. Therefore, our approach could show promising results with patients with disabilities, injected with painkillers, and other categories where physiological scores could not be applied.

Keywords: coma, vigilance state, EEG, multifractal analysis, feature extraction

Procedia PDF Downloads 23

Authors: Tian Lan, Yixiang Wang, Wenxin Tai, Yilan Lyu, Zufeng Wu

Abstract:

The monaural speech enhancement task in the time-frequencydomain has a myriad of approaches, with the stacked con-volutional neural network (CNN) demonstrating superiorability in feature extraction and selection. However, usingstacked single convolutions method limits feature represen-tation capability and generalization ability. In order to solvethe aforementioned problem, we propose an attention-basedadaptive convolutional network that integrates the multi-scale convolutional operations into a operation-specific blockvia input dependent attention to adapt to complex auditoryscenes. In addition, we introduce a two-stage progressivelearning method to enlarge the receptive field without a dra-matic increase in computation burden. We conduct a series ofexperiments based on the TIMIT corpus, and the experimen-tal results prove that our proposed model is better than thestate-of-art models on all metrics.

Keywords: speech enhancement, adaptive convolu-tion, progressive learning, time-frequency domain

Procedia PDF Downloads 89

Authors: Wei Zhang, Yan He, Yan Wang, Yufeng Li, Chuanpeng Hao

Abstract:

Surface roughness is an important index for evaluating surface quality, needs to be accurately measured to ensure the performance of the workpiece. The roughness measurement based on machine vision involves various image features, some of which are redundant. These redundant features affect the accuracy and speed of the visual approach. Previous research used correlation analysis methods to select the appropriate features. However, this feature analysis is independent and cannot fully utilize the information of data. Besides, blindly reducing features lose a lot of useful information, resulting in unreliable results. Therefore, the focus of this paper is on providing a redundant feature removal approach for visual roughness measurement. In this paper, the statistical methods and gray-level co-occurrence matrix(GLCM) are employed to extract the texture features of machined images effectively. Then, the principal component analysis(PCA) is used to fuse all extracted features into a new one, which reduces the feature dimension and maintains the integrity of the original information. Finally, the relationship between new features and roughness is established by the support vector machine(SVM). The experimental results show that the approach can effectively solve multi-feature information redundancy of machined surface images and provides a new idea for the visual evaluation of surface roughness.

Keywords: feature analysis, machine vision, PCA, surface roughness, SVM

Procedia PDF Downloads 184

Authors: Daniel O'Shea, Mario M. Attard, David C. Kellermann

Abstract:

In this paper, we propose a hyperelastic strain energy function that maps isotopic hyperelastic constitutive laws for the use of orthotropic materials without the use of structural tensors or any kind of fiber vector, or the use of standard invariants. In particular, we focus on neo-Hookean class of models and represent them using an invariant-free formulation. To achieve this, we revise the invariant-free formulation of isotropic hyperelasticity. The formulation uses quadruple contractions between fourth-order tensors, rather than scalar products of scalar invariants. We also propose a new decomposition of the orthotropic Hookean stiffness tensor into two fourth-order Lamé tensors that collapse down to the classic Lamé parameters for isotropic continua. The resulting orthotropic hyperelastic model naturally maintains all of the advanced properties of the isotropic counterparts, and similarly collapse back down to their isotropic form by nothing more than equality of parameters in all directions (isotropy). Comparisons are made with large strain experimental results for transversely isotropic rubber type materials under tension.

Keywords: finite strain, hyperelastic, invariants, orthotropic

Procedia PDF Downloads 412

Authors: Yanjun Chen, Chuanping Hu, Jie Shao, Lin Mei, Chongyang Zhang

Abstract:

Traditional image classification is based on encoding scheme (e.g. Fisher Vector, Vector of Locally Aggregated Descriptor) with low-level image features (e.g. SIFT, HoG). Compared to these low-level local features, deep convolutional features obtained at the mid-level layer of convolutional neural networks (CNN) have richer information but lack of geometric invariance. For scene classification, there are scattered objects with different size, category, layout, number and so on. It is crucial to find the distinctive objects in scene as well as their co-occurrence relationship. In this paper, we propose a method to take advantage of both deep convolutional features and the traditional encoding scheme while taking object-centric and scene-centric information into consideration. First, to exploit the object-centric and scene-centric information, two CNNs that trained on ImageNet and Places dataset separately are used as the pre-trained models to extract deep convolutional features at multiple scales. This produces dense local activations. By analyzing the performance of different CNNs at multiple scales, it is found that each CNN works better in different scale ranges. A scale-wise CNN adaption is reasonable since objects in scene are at its own specific scale. Second, a fisher kernel is applied to aggregate a global representation at each scale and then to merge into a single vector by using a post-processing method called scale-wise normalization. The essence of Fisher Vector lies on the accumulation of the first and second order differences. Hence, the scale-wise normalization followed by average pooling would balance the influence of each scale since different amount of features are extracted. Third, the Fisher vector representation based on the deep convolutional features is followed by a linear Supported Vector Machine, which is a simple yet efficient way to classify the scene categories. Experimental results show that the scale-specific feature extraction and normalization with CNNs trained on object-centric and scene-centric datasets can boost the results from 74.03% up to 79.43% on MIT Indoor67 when only two scales are used (compared to results at single scale). The result is comparable to state-of-art performance which proves that the representation can be applied to other visual recognition tasks.

Keywords: deep convolutional features, Fisher Vector, multiple scales, scale-specific normalization

Procedia PDF Downloads 302

Authors: Emel Kızılkaya Aydogan, Mihrimah Ozmen, Yılmaz Delice

Abstract:

In recent days, there is a change and the ongoing development of the telecommunications sector in the global market. In this sector, churn analysis techniques are commonly used for analysing why some customers terminate their service subscriptions prematurely. In addition, customer churn is utmost significant in this sector since it causes to important business loss. Many companies make various researches in order to prevent losses while increasing customer loyalty. Although a large quantity of accumulated data is available in this sector, their usefulness is limited by data quality and relevance. In this paper, a cost-sensitive feature selection framework is developed aiming to obtain the feature reducts to predict customer churn. The framework is a cost based optional pre-processing stage to remove redundant features for churn management. In addition, this cost-based feature selection algorithm is applied in a telecommunication company in Turkey and the results obtained with this algorithm.

Keywords: churn prediction, data mining, decision-theoretic rough set, feature selection

Procedia PDF Downloads 419

Authors: K. P. Sandesh, M. H. Suman

Abstract:

Text processing plays an important role in information retrieval, data-mining, and web search. Measuring the similarity between the documents is an important operation in the text processing field. In this project, a new similarity measure is proposed. To compute the similarity between two documents with respect to a feature the proposed measure takes the following three cases into account: (1) The feature appears in both documents; (2) The feature appears in only one document and; (3) The feature appears in none of the documents. The proposed measure is extended to gauge the similarity between two sets of documents. The effectiveness of our measure is evaluated on several real-world data sets for text classification and clustering problems, especially in banking and health sectors. The results show that the performance obtained by the proposed measure is better than that achieved by the other measures.

Keywords: document classification, document clustering, entropy, accuracy, classifiers, clustering algorithms

Procedia PDF Downloads 480

Authors: Jihan Abdulazeez Ahmed, Adnan Mohsin Abdulazeez Brifcani

Abstract:

This paper suggests a new internal architecture of holon based on feature selection model using the combination of Bees Algorithm (BA) and Artificial Neural Network (ANN). BA is used to generate features while ANN is used as a classifier to evaluate the produced features. Proposed system is applied on the Wine data set, the statistical result proves that the proposed system is effective and has the ability to choose informative features with high accuracy.

Keywords: artificial neural network, bees algorithm, feature selection, Holon

Procedia PDF Downloads 430

Authors: Nougrara Zineb

Abstract:

The unique structure of the blood vessels in the retina has been used for biometric identification. The retina blood vessel pattern is a unique pattern in each individual and it is almost impossible to forge that pattern in a false individual. The retina biometrics’ advantages include high distinctiveness, universality, and stability overtime of the blood vessel pattern. Once the creases have been extracted from the images, a registration stage is necessary, since the position of the retinal vessel structure could change between acquisitions due to the movements of the eye. Image registration consists of following steps: Feature detection, feature matching, transform model estimation and image resembling and transformation. In this paper, we present an algorithm of registration; it is based on the characterization of retinal feature points. For experiments, retinal images from the DRIVE database have been tested. The proposed methodology achieves good results for registration in general.

Keywords: fovea, optic disc, registration, retinal images

Procedia PDF Downloads 236

Authors: Hui Zhang, Ye Tian, Fang Ye, Ziming Guo

Abstract:

Communication signal modulation recognition technology is one of the key technologies in the field of modern information warfare. At present, communication signal automatic modulation recognition methods are mainly divided into two major categories. One is the maximum likelihood hypothesis testing method based on decision theory, the other is a statistical pattern recognition method based on feature extraction. Now, the most commonly used is a statistical pattern recognition method, which includes feature extraction and classifier design. With the increasingly complex electromagnetic environment of communications, how to effectively extract the features of various signals at low signal-to-noise ratio (SNR) is a hot topic for scholars in various countries. To solve this problem, this paper proposes a feature extraction algorithm for the communication signal based on the improved Holder cloud feature. And the extreme learning machine (ELM) is used which aims at the problem of the real-time in the modern warfare to classify the extracted features. The algorithm extracts the digital features of the improved cloud model without deterministic information in a low SNR environment, and uses the improved cloud model to obtain more stable Holder cloud features and the performance of the algorithm is improved. This algorithm addresses the problem that a simple feature extraction algorithm based on Holder coefficient feature is difficult to recognize at low SNR, and it also has a better recognition accuracy. The results of simulations show that the approach in this paper still has a good classification result at low SNR, even when the SNR is -15dB, the recognition accuracy still reaches 76%.

Keywords: communication signal, feature extraction, Holder coefficient, improved cloud model

Procedia PDF Downloads 120

Authors: Han Xue, Zhang Lanyue

Abstract:

In order to seek more effective feature extraction technology, feature extraction method based on MFCC combined with vector hydrophone is exposed in the paper. The sound pressure signal and particle velocity signal of two kinds of ships are extracted by using MFCC and its evolution form, and the extracted features are fused by using fisher-ratio and correlated distance criterion. The features are then identified by BP neural network. The results showed that MFCC, First-Order Differential MFCC and Second-Order Differential MFCC features can be used as effective features for recognition of underwater targets, and the fusion feature can improve the recognition rate. Moreover, the results also showed that the recognition rate of the particle velocity signal is higher than that of the sound pressure signal, and it reflects the superiority of vector signal processing.

Keywords: vector information, MFCC, differential MFCC, fusion feature, BP neural network

Procedia PDF Downloads 491

Authors: Souvik Phadikar, Nidul Sinha, Rajdeep Ghosh

Abstract:

In brainwaves research using electroencephalogram (EEG) signals, finding the most relevant and effective feature set for identification of activities in the human brain is a big challenge till today because of the random nature of the signals. The feature extraction method is a key issue to solve this problem. Finding those features that prove to give distinctive pictures for different activities and similar for the same activities is very difficult, especially for the number of activities. The performance of a classifier accuracy depends on this quality of feature set. Further, more number of features result in high computational complexity and less number of features compromise with the lower performance. In this paper, a novel idea of the selection of optimal feature set using a heuristically optimized deep autoencoder is presented. Using various feature extraction methods, a vast number of features are extracted from the EEG signals and fed to the autoencoder deep neural network. The autoencoder encodes the input features into a small set of codes. To avoid the gradient vanish problem and normalization of the dataset, a meta-heuristic search algorithm is used to minimize the mean square error (MSE) between encoder input and decoder output. To reduce the feature set into a smaller one, 4 hidden layers are considered in the autoencoder network; hence it is called Heuristically Optimized Deep Autoencoder (HO-DAE). In this method, no features are rejected; all the features are combined into the response of responses of the hidden layer. The results reveal that higher accuracy can be achieved using optimal reduced features. The proposed HO-DAE is also compared with the regular autoencoder to test the performance of both. The performance of the proposed method is validated and compared with the other two methods recently reported in the literature, which reveals that the proposed method is far better than the other two methods in terms of classification accuracy.

Keywords: autoencoder, brainwave signal analysis, electroencephalogram, feature extraction, feature selection, optimization

Procedia PDF Downloads 92

Authors: Bubai Maji, Monorama Swain

Abstract:

Speech is considered an essential and most natural medium for the interaction between machines and humans. However, extracting effective features for speech emotion recognition (SER) is remains challenging. The present studies show that the temporal information captured but high-level temporal-feature learning is yet to be investigated. In this paper, we present an efficient novel method using the Self-attention (SA) mechanism in a combination of Convolutional Neural Network (CNN) and Bi-directional Gated Recurrent Unit (Bi-GRU) network to learn high-level temporal-feature. In order to further enhance the representation of the high-level temporal-feature, we integrate a Bi-GRU output with learnable weights features by SA, and improve the performance. We evaluate our proposed method on our created SITB-OSED and IEMOCAP databases. We report that the experimental results of our proposed method achieve state-of-the-art performance on both databases.

Keywords: Bi-GRU, 1D-CNNs, self-attention, speech emotion recognition

Procedia PDF Downloads 90

Authors: Jingting Luo, Yong Wang, Ying Wang

Abstract:

Pedestrian detection in image or video data is a very important and challenging task in security surveillance. The difficulty of this task is to locate and detect pedestrians of different scales in complex scenes accurately. To solve these problems, a deep neural network (RT-YOLOv3) is proposed to realize real-time pedestrian detection at different scales in security monitoring. RT-YOLOv3 improves the traditional YOLOv3 algorithm. Firstly, the deep residual network is added to extract vehicle features. Then six convolutional neural networks with different scales are designed and fused with the corresponding scale feature maps in the residual network to form the final feature pyramid to perform pedestrian detection tasks. This method can better characterize pedestrians. In order to further improve the accuracy and generalization ability of the model, a hybrid pedestrian data set training method is used to extract pedestrian data from the VOC data set and train with the INRIA pedestrian data set. Experiments show that the proposed RT-YOLOv3 method achieves 93.57% accuracy of mAP (mean average precision) and 46.52f/s (number of frames per second). In terms of accuracy, RT-YOLOv3 performs better than Fast R-CNN, Faster R-CNN, YOLO, SSD, YOLOv2, and YOLOv3. This method reduces the missed detection rate and false detection rate, improves the positioning accuracy, and meets the requirements of real-time detection of pedestrian objects.

Keywords: pedestrian detection, feature detection, convolutional neural network, real-time detection, YOLOv3

Procedia PDF Downloads 115