Showing papers in "Signal, Image and Video Processing in 2017"
TL;DR: A novel passive image forgery detection method is proposed based on local binary pattern (LBP) and discrete cosine transform (DCT) to detect copy–move and splicing forgeries.
Abstract: With the development of easy-to-use and sophisticated image editing software, the alteration of the contents of digital images has become very easy to do and hard to detect. A digital image is a very rich source of information and can capture any event perfectly, but because of this reason, its authenticity is questionable. In this paper, a novel passive image forgery detection method is proposed based on local binary pattern (LBP) and discrete cosine transform (DCT) to detect copy–move and splicing forgeries. First, from the chrominance component of the input image, discriminative localized features are extracted by applying 2D DCT in LBP space. Then, support vector machine is used for detection. Experiments carried out on three image forgery benchmark datasets demonstrate the superiority of the method over recent methods in terms of detection accuracy.
99 citations
TL;DR: In this study, a new S-box construction method based on fractional-order (FO) chaotic Chen system is presented and it is observed that this method provides a stronger S- box design.
Abstract: Since substitution box (S-box) is the only nonlinear component related to confusion properties for many block encryption algorithms, it is a necessity for the strong block encryption algorithms. S-box is a vital component in cryptography due to having the effect on the security of entire system. Therefore, alternative S-box construction techniques have been proposed in many researches. In this study, a new S-box construction method based on fractional-order (FO) chaotic Chen system is presented. In order to achieve that goal, numerical results of the FO chaotic Chen system for $$a= 35, b=3, c=28$$
and $$\alpha =0.9$$
are obtained by employing the predictor–corrector scheme. Besides, a simpler algorithm is suggested for the construction of S-box via time response of the FO chaotic Chen system. The performance of suggested S-box design is compared with other S-box designs developed by chaotic systems, and it is observed that this method provides a stronger S-box design.
99 citations
TL;DR: This paper presents an efficient and non-intrusive method to counter face-spoofing attacks that uses a single image to detect spoofing attacks and proposes a specialized deep convolution neural network that can extract the discriminative and high-level features of the input diffused image to differentiate between a fake face and a real face.
Abstract: A face-spoofing attack occurs when an imposter manipulates a face recognition and verification system to gain access as a legitimate user by presenting a 2D printed image or recorded video to the face sensor. This paper presents an efficient and non-intrusive method to counter face-spoofing attacks that uses a single image to detect spoofing attacks. We apply a nonlinear diffusion based on an additive operator splitting scheme. Additionally, we propose a specialized deep convolution neural network that can extract the discriminative and high-level features of the input diffused image to differentiate between a fake face and a real face. Our proposed method is both efficient and convenient compared with the previously implemented state-of-the-art methods described in the literature review. We achieved the highest reported accuracy of 99% on the widely used NUAA dataset. In addition, we tested our method on the Replay Attack dataset which consists of 1200 short videos of both real access and spoofing attacks. An extensive experimental analysis was conducted that demonstrated better results when compared to previous static algorithms results. However, this result can be improved by applying a sparse autoencoder learning algorithm to obtain a more distinguishable diffused image.
95 citations
TL;DR: A new approach to detect fire from a video stream that takes full advantage of the motion feature and color information of fire to achieve better effectiveness, adaptability and robustness.
Abstract: This paper proposes a new approach to detect fire from a video stream. It takes full advantage of the motion feature and color information of fire. Firstly, motion detection using Gaussian Mixture Model-based background subtraction is applied to extract moving objects from a video stream. Then, multi-color-based detection combining the RGB, HSI and YUV color space is employed to obtain possible fire regions. Finally, the results of the above two steps are combined to identify the accurate fire areas. The experimental results obtained by applying this method on different fire videos show that the proposed method can achieve better effectiveness, adaptability and robustness.
73 citations
TL;DR: An approach to segment the moving objects using both the frame differencing and W4 algorithm to overcome the above problems and the effectiveness of this approach in comparison with existing techniques is demonstrated.
Abstract: Moving object detection is a basic and important task on automated video surveillance systems, because it gives the focus of attention for further examination. Frame differencing and W4 algorithm can be individually employed to detect the moving objects. However, the detected results of the individual approach are not accurate due to foreground aperture and ghosting problems. We propose an approach to segment the moving objects using both the frame differencing and W4 algorithm to overcome the above problems. Here first we compute the difference between consecutive frames using histogram-based frame differencing technique, next W4 algorithm is applied on frame sequences, and subsequently, the outcomes of the frame differencing and W4 algorithm are combined using logical ‘OR’ operation. Finally, morphological operation with connected component labeling is employed to detect the moving objects. The experimental results and performance evaluation on real video datasets demonstrate the effectiveness of our approach in comparison with existing techniques.
65 citations
TL;DR: A deep learning-based anomaly detection system that achieves complete detection of abnormal events by involving the following significant proposed modules a Background Estimation Module, an Object Segmentation Module, a Feature Extraction Module, and an Activity Recognition (AR) Module.
Abstract: In this paper, a deep learning-based anomaly detection (DLAD) system is proposed to improve the recognition problem in video processing. Our system achieves complete detection of abnormal events by involving the following significant proposed modules a Background Estimation (BE) Module, an Object Segmentation (OS) Module, a Feature Extraction (FE) Module, and an Activity Recognition (AR) Module. At first, we have presented a BE (Background Estimation) module that generated an accurate background in which two-phase model is generated to compute the background estimation. After a high-quality background is generated, the OS model is developed to extract the object from videos, and then, object tracking process is used to track the object through the overlapping detection scheme. From the tracked objects, the FE module is extracted for some useful features such as shape, wavelet, and histogram to the abnormal event detection. For the final step, the proposed AR module is classified as abnormal or normal event using the deep learning classifier. Experiments are performed on the USCD benchmark dataset of abnormal activities, and comparisons with the state-of-the-art methods validate the advantages of our algorithm. We can see that the proposed activity recognition system has outperformed by achieving better EER of 0.75 % when compared with the existing systems (20 %). Also, it shows that the proposed method achieves 85 % precision rate in the frame-level performance.
61 citations
TL;DR: A new approach for breast thermogram image analysis is presented by developing a fully automatic segmentation of right and left breast for asymmetry analysis, using shape features of the breast and Polynomial curve fitting.
Abstract: In this article, we present a new approach for breast thermogram image analysis by developing a fully automatic segmentation of right and left breast for asymmetry analysis, using shape features of the breast and Polynomial curve fitting. Segmentation results are validated with their respective Ground Truths. Histogram and grey level co-occurrence matrix-based texture features are extracted from the segmented images. Statistical test shows that features are highly significant in detection of breast cancer. We have obtained an accuracy of 90%, sensitivity of 87.5% and specificity of 92.5% for a set of eighty images with forty normal and forty abnormal using SVM RBF classifier.
57 citations
TL;DR: A novel software-based fingerprint liveness detection method which achieves good detection accuracy and outperform the state-of-the-art methods is proposed.
Abstract: Fingerprint-based recognition systems have been increasingly deployed in various applications nowadays. However, the recognition systems can be spoofed by using an accurate imitation of a live fingerprint such as an artificially made fingerprint. In this paper, we propose a novel software-based fingerprint liveness detection method which achieves good detection accuracy. We regard the fingerprint liveness detection as a two-class classification problem and construct co-occurrence array from image gradients to extract features. In doing so, the quantization operation is firstly conducted on the images. Then, the horizontal and vertical gradients at each pixel are calculated, and the gradients of large absolute values are truncated into a reduced range. Finally, the second-order and the third-order co-occurrence arrays are constructed from the truncated gradients, and the elements of the co-occurrence arrays are directly used as features. The second-order and the third-order co-occurrence array features are separately utilized to train support vector machine classifiers on two publicly available databases used in Fingerprint Liveness Detection Competition 2009 and 2011. The experimental results have demonstrated that the features extracted with the third-order co-occurrence array achieve better detection accuracy than that with the second-order co-occurrence array and outperform the state-of-the-art methods.
52 citations
TL;DR: A new segmentation method is proposed to address the issue of low sensitivity, by including modules such as principal component analysis-based color-to-gray conversion, scale normalization factors for improved narrow vessel detection, anisotropic diffusion filtering with an adequate stopping rule, and edge pixel-based hysteresis threshold.
Abstract: Retinal vessel segmentation plays a major role in the detection of many eye diseases, and it is required to implement an automated algorithm for analyzing the progress of eye diseases. A variety of automated segmentation methods have been presented but almost all studies to date showed weakness in their low sensitivity toward narrow low-contrast vessels. A new segmentation method is proposed to address the issue of low sensitivity, by including modules such as principal component analysis-based color-to-gray conversion, scale normalization factors for improved narrow vessel detection, anisotropic diffusion filtering with an adequate stopping rule, and edge pixel-based hysteresis threshold. The impact of these additional steps is assessed on publicly available databases like DRIVE and STARE. For the case of DRIVE database, the sensitivity is raised from 73 to 75%, while maintaining the accuracy of 96.5%, and found to provide evidence of improved sensitivity.
41 citations
TL;DR: The effectiveness of the proposed grayscale conversion is confirmed by the comparative analysis performed on the color-to-gray benchmark dataset across 10 existing algorithms based on the standard objective measures, namely normalized cross-correlation, color contrast preservation ratio, color content fidelity ratio, E score and subjective evaluation.
Abstract: This paper provides an alternative framework for color-to-grayscale image conversion by exploiting the chrominance information present in the color image using singular value decomposition (SVD). In the proposed technique of color-to-grayscale image conversion, a weight matrix corresponds to the chrominance components is derived by reconstructing the chrominance data matrix (planes a* and b*) from the eigenvalues and eigenvectors computed using SVD. The final grayscale converted image is obtained by adding the weighted chrominance data to the luminous intensity which is kept intact for the CIEL*a*b* color space of the given color image. The effectiveness of the proposed grayscale conversion is confirmed by the comparative analysis performed on the color-to-gray benchmark dataset across 10 existing algorithms based on the standard objective measures, namely normalized cross-correlation, color contrast preservation ratio, color content fidelity ratio, E score and subjective evaluation.
41 citations
TL;DR: A fast context-sensitive threshold selection technique that incorporates spatial contextual information of the image in threshold selection process without loosing the benefits of histogram-based techniques is presented to solve the image segmentation problems.
Abstract: In this article, a fast context-sensitive threshold selection technique is presented to solve the image segmentation problems. In lieu of histogram, the proposed technique employs recently defined energy curve of the image. First, the initial thresholds are selected in the middle of two consecutive peaks on the energy curve. Then based on the cluster validity measure, the optimal number of potential thresholds and the bounds where the optimal value of each potential threshold may exist are determined. Finally, genetic algorithm (GA) is employed to detect the optimal value of each potential threshold from their respective defined bounds. The proposed technique incorporates spatial contextual information of the image in threshold selection process without loosing the benefits of histogram-based techniques. Computationally it is very efficient. Moreover, it is able to determine the optimal number of segments in the input image. To assess the effectiveness of the proposed technique, the results obtained are compared with four state-of-the-art methods cited in the literature. Experimental results on large number of images confirmed the effectiveness of the proposed technique.
TL;DR: The result of the method indicates its robustness for glaucoma evaluation and incorporates masking to avoid misclassifying areas as well as forming the structure of the cup based on edge detection.
Abstract: This research proposes a robust method for disc localization and cup segmentation that incorporates masking to avoid misclassifying areas as well as forming the structure of the cup based on edge detection. Our method has been evaluated using two fundus image datasets, namely: D-I and D-II comprising of 60 and 38 images, respectively. The proposed method of disc localization achieves an average $$F_{\mathrm{score}}$$
of 0.96 and average boundary distance of 7.7 for D-I, and 0.96 and 9.1, respectively, for D-II. The cup segmentation method attains an average $$F_{\mathrm{score}}$$
of 0.88 and average boundary distance of 13.8 for D-I, and 0.85 and 18.0, respectively, for D-II. The estimation errors (mean ± standard deviation) of our method for the value of vertical cup-to-disc diameter ratio against the result of the boundary by the expert of D-I and D-II have similar value, namely $$0.04 \pm 0.04$$
. Overall, the result of our method indicates its robustness for glaucoma evaluation.
TL;DR: A novel deformable model is proposed for efficient 3D visual tracking of beating heart and is validated on the stereo-endoscopic videos of phantom heart and in vivo heart that are recorded by the da Vinci surgical system.
Abstract: A novel deformable model is proposed for efficient 3D visual tracking of beating heart. The model is parameterized by the 3D coordinates of four control points: the three vertices and the circumcenter of a triangular target region. Nonlinear deformation on heart surfaces is handled by cubic spline interpolation based on radial pixel distances from the circumcenter. With a pre-computable design matrix, the model can be represented efficiently by a simple matrix equation. An iterative algorithm is developed based on the efficient second-order minimization to compute model parameters at each frame. The proposed tracking method is validated on the stereo-endoscopic videos of phantom heart and in vivo heart that are recorded by the da Vinci
$$^{\tiny \textregistered }$$
surgical system.
TL;DR: The proposed supervised feature extraction approach that is capable of selecting distinctive features for the recognition of human gait under clothing and carrying conditions, thus improving the recognition performances, is proposed.
Abstract: This paper proposes a supervised feature extraction approach that is capable of selecting distinctive features for the recognition of human gait under clothing and carrying conditions, thus improving the recognition performances. The principle of the suggested approach is based on the Haralick features extracted from gait energy image (GEI). These features are extracted locally by dividing vertically or horizontally the GEI locally into two or three equal regions of interest, respectively. RELIEF feature selection algorithm is then employed on the extracted features in order to select only the most relevant features with a minimum redundancy. The proposed method is evaluated on CASIA gait database (Dataset B) under variations of clothing and carrying conditions for different viewing angles, and the experimental results using k-NN classifier have yielded attractive results of up to 80% in terms of highest identification rate at rank-1 when compared to existing and similar state-of-the-art methods.
TL;DR: This paper presents a novel solution for distinguishing between live and forged identities using the fusion of texture-based methods and image quality assessment measures using LBP and HOG texture descriptors to extract texture information of an image.
Abstract: Spoofing attacks made by non-real images are a major concern to biometric systems. This paper presents a novel solution for distinguishing between live and forged identities using the fusion of texture-based methods and image quality assessment measures. In our approach, we used LBP and HOG texture descriptors to extract texture information of an image. Additionally, feature space of seven full-reference complementary image quality measures is considered including peak signal-to-noise ratio, structural similarity, mean-squared error, normalized cross-correlation, maximum difference, normalized absolute error and average difference. We built a palmprint spoof database made by printed palmprint photograph of PolyU palmprint database using camera. Experimental results on three public-domain face spoof databases (Idiap Print-Attack, Replay-Attack and MSU MFSD) and palmprint spoof database show that the proposed solution is effective in face and palmprint spoof detection.
TL;DR: Experimental results indicate that the BHE2PL method exhibits a better mean brightness preservation compared to methods found in the state of the art; in addition to also presenting a reasonable computation time.
Abstract: Histogram equalization is an effective method for contrast enhancement on images, but it suffers from some problems such as the tendency to change the mean brightness, loss of information and the introduction of saturation levels which causes an unnatural appearance in the resulting image. Due to the aforementioned problems, a variety of histogram equalization methods have been developed in order to preserve the image brightness, thus avoiding saturation levels that cause loss of information. In this paper, the bi-histogram equalization using two plateau limits (BHE2PL) for histogram equalization is proposed. BHE2PL divides the global histogram into two sub-histograms; then, each sub-histogram is modified by two plateau limits in order to avoid over-enhancement of the image. Experimental results indicate that the BHE2PL method exhibits a better mean brightness preservation compared to methods found in the state of the art; in addition to also presenting a reasonable computation time.
TL;DR: Experimental results show that the proposed improved method using Kernel Regression Model on local neighbor data can address the problems of being unsmooth and the absence of neighbor information for the transmission estimation under Dark Channel Prior framework and has better dehazing results than several state-of-the-art methods.
Abstract: Haze is one of the major factors that degrade outdoor images, and dehazing becomes an important issue in many applications. In order to address the problems of being unsmooth and the absence of neighbor information for the transmission estimation under Dark Channel Prior (DCP) framework, we proposed a new improved method using Kernel Regression Model (KRM) on local neighbor data. Firstly, the initial transmission in atmospheric light model is estimated by DCP. Secondly, the transmission is refined according to KRM. Experimental results on synthetic and real images show that our method can address this problem and has better dehazing results than several state-of-the-art methods.
TL;DR: A computational algorithm based on hybrid model to automatically extract vision perception features from raw image patches is proposed, which demonstrates very competitive quality prediction performance of the proposed method.
Abstract: The aim of research on the no-reference image quality assessment problem is to design models that can predict the quality of distorted images consistently with human visual perception. Due to the little prior knowledge of the images, it is still a difficult problem. This paper proposes a computational algorithm based on hybrid model to automatically extract vision perception features from raw image patches. Convolutional neural network (CNN) and support vector regression (SVR) are combined for this purpose. In the hybrid model, the CNN is trained as an efficient feature extractor, and the SVR performs as the regression operator. Extensive experiments demonstrate very competitive quality prediction performance of the proposed method.
TL;DR: A new algorithm for classification of ground vehicles from standard synthetic aperture radar (SAR) images is proposed to use RCM as the feature extraction mechanism and to compare results of the fused images with both Zernike and radial Chebyshev moments.
Abstract: In this study, a new algorithm for classification of ground vehicles from standard synthetic aperture radar (SAR) images is proposed. Radial Chebyshev moment (RCM) is a discrete orthogonal moment that has distinctive advantages over other moments for feature extraction. Unlike invariant moments, its orthogonal basis leads to having minimum information redundancy, and its discrete characteristics explore some benefits over Zernike moments (ZM) due to having no numerical errors and no computational complexity owing to normalization. In this context, we propose to use RCM as the feature extraction mechanism on the segmented image and to compare results of the fused images with both Zernike and radial Chebyshev moments. Firstly, by applying different threshold target and shadow parts of each SAR images are extracted separately. Then, segmented images are fused based on the combination of the extracted segmented region, segmented boundary and segmented texture. Experimental results will verify that accuracy of RCM, which improves significantly over the ZM. Ten percent improvement in the accuracy is obtained by using RCM and fusion of segmented target and shadow parts. Furthermore, feature fusion improves the total accuracy of the classification as high as 6%.
TL;DR: This work performs dimensionality reduction-based classification on fleece fabric-based images taken by a thermal camera using Naive Bayes and K-nearest neighbor classifier with great classification accuracy.
Abstract: This work performs dimensionality reduction-based classification on fleece fabric-based images taken by a thermal camera. In order to convert images into the gray level, a principal component analysis-based dimension reduction stage was proposed. In addition, symmetric central local binary patterns were performed with the help of the proposed method by using the images after dimension reduction process. The local binary pattern features preserve local texture features from different kinds of defective image types. The experimental results showed that combined work has a great classification accuracy. The classification accuracy was reported using two different algorithms: Naive Bayes and K-nearest neighbor classifier.
TL;DR: A novel multi-focus image fusion technique is presented, developed by using the nonsubsampled contourlet transform (NSCT) and a proposed fuzzy logic based adaptive pulse-coupled neural network (PCNN) model, where sum-modified Laplacian (SML) is calculated as the motivation for PCNN neurons in NSCT domain.
Abstract: Multi-focus image fusion technique can solve the problem that not all the targets in an image are clear in case of imaging in the same scene. In this paper, a novel multi-focus image fusion technique is presented, which is developed by using the nonsubsampled contourlet transform (NSCT) and a proposed fuzzy logic based adaptive pulse-coupled neural network (PCNN) model. In our method, sum-modified Laplacian (SML) is calculated as the motivation for PCNN neurons in NSCT domain. Since the linking strength plays an important role in PCNN, we propose an adaptively fuzzy way to determine it by computing each coefficient’s importance relative to the surrounding coefficients. Combined with human visual perception characteristics, the fuzzy membership value is employed to automatically achieve the degree of importance of each coefficient, which is utilized as the linking strength in PCNN model. Experimental results on simulated and real multi-focus images show that the proposed technique has a superior performance to series of exist fusion methods.
TL;DR: An ANFIS technique to predict the friction coefficient as output variable based on pipe relative roughness and Reynold's number as input variables is developed and it was found that the adaptive neuro-fuzzy inference system model is more accurate than other empirical equations in modeling friction factor.
Abstract: The friction coefficient is widely used for technical and economical design of pipes in irrigation, land drainage, urban sewage systems and intake structures. In the present study, the friction factor in pipes is estimated by using adaptive neuro-fuzzy inference system (ANFIS) and grid partition method. The data derived from the Colebrook’s equation were considered for ascertaining the neuro-fuzzy model. Present approach developed an ANFIS technique to predict the friction coefficient as output variable based on pipe relative roughness and Reynold’s number as input variables. The performance of the ANFIS model was evaluated against conventional procedures. Correlation coefficient (R2), root mean squared error and mean absolute error were used as comparing statistical indicators for the assessment of the proposed approach’s performance. It was found that the adaptive neuro-fuzzy inference system model is more accurate than other empirical equations in modeling friction factor.
TL;DR: Experimental results indicate that the proposed fast super-resolution method based on first-order derivatives from neighbor pixels clearly outperforms other state-of-the-art algorithms such as fast curvature-based interpolation.
Abstract: A single fast super-resolution method based on first-order derivatives from neighbor pixels is proposed. The basic idea of the proposed method is to exploit a first-order derivatives component of six edge directions around a missing pixel, followed by back projection to reduce noise estimated by the difference between simulated and observed images. Using first-order derivatives as a feature, the proposed method is expected to have low computational complexity, and it can theoretically reduce blur, blocking, and ringing artifacts in edge areas compared to previous methods. Experiments were conducted using 900 natural grayscale images from the USC-SIPI Database. We evaluated the proposed and previous methods using peak signal-to-noise ratio, structural similarity, feature similarity, and computation time. Experimental results indicate that the proposed method clearly outperforms other state-of-the-art algorithms such as fast curvature-based interpolation.
TL;DR: This paper presents a fast method for moving object detection that improves the existing methods by updating the Gaussian mixture model selectively, and shows that computation time of the proposed method is reduced.
Abstract: Moving object detection and extraction are widely used in video surveillance and image processing. In this paper, we present a fast method for moving object detection. We use weights of the Gaussian distribution as decision factors, update parameters of the Gaussian mixture model if its values are smaller than that of those not belonging to the background; otherwise, no updates are done. It improves the existing methods by updating the Gaussian mixture model selectively. Experimental results on various scenes of video surveillance show that computation time of the proposed method is reduced.
TL;DR: The particle swarm optimization algorithm performs the image registration using the generated control and trajectory points, observed in both the cameras, and demonstrates the advantages of fusing the thermal and visible camera within a pedestrian detection algorithm.
Abstract: In this paper, we propose methods to calibrate visible and thermal cameras and register their images in the application of pedestrian detection. We calibrate the camera using a checkerboard pattern mounted on a heated rig. We implement the image registration using three different approaches. In the first approach, we use the camera calibration information to generate control points from the checkerboard pattern. These control points are then used to register the images. In the second approach, we generate trajectory points for image registration using an external illuminated object. In the third approach, we achieve the registration through face tracking without the aid of any external object. The particle swarm optimization algorithm performs the image registration using the generated control and trajectory points, observed in both the cameras. We demonstrate the advantages of fusing the thermal and visible camera within a pedestrian detection algorithm. We evaluate the proposed registration algorithms and perform a comparison with baseline algorithms, i.e. genetic and simulated annealing algorithms. Additionally, we also perform a detailed parameter evaluation of the particle swarm optimization algorithm. The experimental results demonstrate the accuracy of the proposed algorithm and the advantages of thermal-visible camera fusion.
TL;DR: Experimental results have revealed that the proposed patch selection strategy, based on the regions of interest, can improve quality measures of three IQA methods.
Abstract: Most methods in the literature of image quality assessment (IQA) use whole image information for measuring image quality. However, human perception does not always use this criterion to assess the quality of images. Individuals usually provide their opinions by considering only some parts of an image, called regions of interest. Based on this hypothesis, in this research work, a segmentation technique is initially employed to obtain a bi-level image map composed of the foreground and background information. A patch selection strategy is then proposed to choose some particular patches based on the foreground information as the regions of interest for IQA. Three recent IQA methods in the literature are considered to demonstrate the improvement in IQA when using only the extracted regions of interest. To evaluate the impact of the proposed patch selection strategy in various IQA metrics, three publicly available datasets were used for experiments. Experimental results have revealed that our proposal, based on the regions of interest, can improve quality measures of three IQA methods.
TL;DR: Radiance profiles of burn scar are generated for the observed atmospheric and illumination conditions at the time of the hyperspectral image data collection and form a radiance profile library using a nonlinear analytical model for radiative transfer and MODTRAN.
Abstract: Assessment of damages due to fire, drought, flood, land slide, etc., using hyperspectral images from Hyperion, AVIRIS or HyspIRI has challenging issues. The effects of different illumination, atmospheric conditions and varying sensor/target viewing geometries are some of these challenges. A common approach for target detection is to apply atmospheric correction algorithms to the radiance image data cube and then search within the atmospherically corrected image cube for the target reflectance signature of interest. One major issue with the above approach is that it is computationally demanding. In this paper, instead of applying atmospheric correction to the raw radiance data, we generate radiance profiles of burn scar for the observed atmospheric and illumination conditions at the time of the hyperspectral image data collection and form a radiance profile library using a nonlinear analytical model for radiative transfer and MODTRAN. The target detection has been performed by a spectral similarity technique which takes into consideration multiple radiance profile variants of the target of interest. The effectiveness of the radiance domain-based target detection approach on reducing the computation time has been demonstrated on burn scar detection using airborne AVIRIS image data.
TL;DR: A multilevel classification approach based on single-branch decision tree has been proposed for improved facial expression recognition and outperform most of its counterparts in the literature under the same databases and settings.
Abstract: In this paper, a new approach has been proposed for improved facial expression recognition. The new approach is inspired by the compressive sensing theory and multiresolution approach to facial expression problems. Initially, each image sample is decomposed into desired pyramid levels at different sizes and resolutions. Pyramid features at all levels are concatenated to form a pyramid feature vector. The vectors are further reinforced and reduced in dimension using a measurement matrix based on compressive sensing theory. For classification, a multilevel classification approach based on single-branch decision tree has been proposed. The proposed multilevel classification approach trains a number of binary support vector machines equal to the number of classes in the datasets. Class of test data is evaluated through the nodes of the tree from the root to its apex. The results obtained from the approach are impressive and outperform most of its counterparts in the literature under the same databases and settings.
TL;DR: Empirical studies for practical applications including robust principal component analysis and low-rank representation demonstrate that the proposed algorithm outperforms many other state-of-the-art convex and non-convex methods developed recently in the literature.
Abstract: This paper concerns the low-rank minimization problems which consist of finding a matrix of minimum rank subject to linear constraints. Many existing approaches, which used the nuclear norm as a convex surrogate of the rank function, usually result in a suboptimal solution. To seek a tighter rank approximation, we develop a non-convex surrogate to approximate the rank function based on the Laplace function. An iterative algorithm based on the augmented Lagrangian multipliers method is developed. Empirical studies for practical applications including robust principal component analysis and low-rank representation demonstrate that our proposed algorithm outperforms many other state-of-the-art convex and non-convex methods developed recently in the literature.
TL;DR: This paper presents a new fast edge-based stereo matching approach devoted to road applications that has been tested on both real and virtual stereo images, it has been compared to a recently proposed method, and the results are satisfactory.
Abstract: Several vision-based road applications use stereo vision algorithms, and they generally must be fast to be applied in real time. The main problem in stereo vision is the stereo matching problem, which consists in finding correspondences between two stereo images. In this paper, we present a new fast edge-based stereo matching approach devoted to road applications. Two passes of the dynamic programming algorithm are applied to estimate the final disparity map. The matching results of the first pass are only exploited to compute an initial disparity map (IDM). The so-called guiding edge points (GEPs) together with disparity ranges, i.e., possible matches, are derived from the IDM. In the second pass, the disparity ranges are used to reduce the search space as well as the mismatches and the GEPs to control and guide the matching process to the optimal solution. The proposed method has been tested on both real and virtual stereo images, it has been compared to a recently proposed method, and the results are satisfactory.