Journal of Visual Communication and Image Representation, Feb 1, 2019
In this work, we present a fully connected convolutional encoder-decoder for defects detection in... more In this work, we present a fully connected convolutional encoder-decoder for defects detection in archived video. The proposed method handles the detection of two of the most common archived video-related defects, namely blotches and scratches. It consists of two stages: 1) pixel-level classification and description of each video frame into defects pixels or not, by means of a novel CNN-based encoder-decoder architecture, and 2) spatio-temporal analysis to group and fine-tune the detections. For blotch detection, the learned features, extracted from an intermediate stage of the network, are used to evaluate the dissimilarity between the pre-selected regions in consecutive frames. For scratches detection, the morphology of scratches is used to eliminate false alarms. The experiments are performed on various video sequences suffering synthetic and real scratches and blotches. The results demonstrate the effectiveness of our approach and significant improvement against the most recent detectors.
Abstract. We propose a method for blotch detection in archived videos by modeling their spatiotem... more Abstract. We propose a method for blotch detection in archived videos by modeling their spatiotemporal properties. We introduce an adaptive spatiotemporal segmentation to extract candidate regions that can be classified as blotches. Then, the similarity between the preselected regions and their corresponding motion-compensated regions in the adjacent frames is assessed by means of motion trajectory estimation and textural information analysis. Perceived ground truth based on just noticeable contrast is employed for the evaluation of our approach against the state-of-the-art, and the reported results show a better performance for our approach.
2017 European Conference on Electrical Engineering and Computer Science (EECS)
Image segmentation is often required as a fundamental stage in microstructure material characteri... more Image segmentation is often required as a fundamental stage in microstructure material characterization. The objective of this work is to choose hybridization between the Level Set method and the clustering approach in order to extract the characteristics of the materials from the segmentation result of the micrographic images. More specifically, the proposed approach contains two successive necessary stages. The first one consists in the application of possibilistic c-means clustering approach (PCM) to get the various classes of the original image. The second stage is based on using the result of the clustering approach i.e. one class among the three existing classes (which interests us) as an initial contour of the level set method to extract the boundaries of interest region. The main purpose of using the result of the PCM algorithm as initial step of the level set method is to enhance and facilitate the work of the latter. Our experimental results on real micrographic images show that the proposed segmentation method can extract successfully the interest region according to the chosen class and confirm its efficiency for segmenting micrographic images of materials.
This paper turns on blotches detection in archived video. Blotches could be considered as the mos... more This paper turns on blotches detection in archived video. Blotches could be considered as the most common artifacts in archived video. This type of artifacts is not temporally correlated, it can be considered as impulse noise in time. To detect blotches, the proposed approach is performed into two mains steps: blotch regions candidate extraction and classification of these regions as blotches or not. The first step is performed by using sudden changes region detection by means of median filtering in time. The second step relies on finding similarity of the candidate region in the adjacent frames in gradient space. The proposed method has been compared against ROD, SROD and SDIx detectors. The obtained results show that our proposed approach outperforms the considered methods.
2017 International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)
Verification of individual identity through the process of biometric identification involves comp... more Verification of individual identity through the process of biometric identification involves comparison between an encoded value and a stored value of the biometric feature in question. The effectiveness of a multimodal user authentication system is greater, but so is its complexity. The system error rate is reduced by the fact that multiple biometric features are combined, thus solving the weakness of the single biometric. Performance of individual authentication through palm-print-and iris-based bimodal biometric system is proposed in the present study. To this end, Log-Gabor filter and BSIF (Binarised Statistical Image Feature) coefficients are employed to obtain the iris and palmprint traits, and subsequently selection of the features vector is conducted with mRMR (Minimum Redundancy Maximum Relevance) transforms in higher coefficients. To match the iris or palm-print feature vector, the Hamming Distance is applied. According to the experiment outcomes, the proposed system not only has a significantly high recognition rate but it also affords greater security compared to the single biometric system.
2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET), 2017
In Automatic Fingerprint Identification System (AFIS), the performance highly depends on features... more In Automatic Fingerprint Identification System (AFIS), the performance highly depends on features extraction. In this paper, a new minutiae detection method based on wave atoms transform is presented. The proposed approach is based on analyzing fingerprints by specific representation highlighting the oscillatory structure. Thereby, a fine analysis from local characteristics via the block wave atoms transform is introduced. We explain how minutiae are detected by the use of wave atoms transform associated to a support vector machine (SVM) classifier. In order to evaluate the algorithm, FVC 2002 fingerprint databases have been considered. The tests show that for fingerprints features extraction, the wave atoms transform offers a good performance.
2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP), 2020
This paper deals with the detection of blurred images that may eventually be captured by a drone.... more This paper deals with the detection of blurred images that may eventually be captured by a drone. The embedded system should be able to measure the amount of blur affecting the images in order to decide whether to acquire the scene again or not. For this purpose, we have developed a simple model based on Discrete Cosine Transform (DCT) associated to Support Vector Machine Classifier SVM, to classify images into three categories and thus detect strongly, moderately and slightly blurred images. The proposed system has been tested on 550 images captured by a drone. The obtained results are very conclusive.
Palmprint recognition systems are dependent on feature extraction. A method of feature extraction... more Palmprint recognition systems are dependent on feature extraction. A method of feature extraction using higher discrimination information was developed to characterize palmprint images. In this method, two individual feature extraction techniques are applied to a discrete wavelet transform of a palmprint image, and their outputs are fused. The two techniques used in the fusion are the histogram of gradient and the binarized statistical image features. They are then evaluated using an extreme learning machine classifier before selecting a feature based on principal component analysis. Three palmprint databases,
Journal of Visual Communication and Image Representation, 2019
In this work, we present a fully connected convolutional encoder-decoder for defects detection in... more In this work, we present a fully connected convolutional encoder-decoder for defects detection in archived video. The proposed method handles the detection of two of the most common archived video-related defects, namely blotches and scratches. It consists of two stages: 1) pixel-level classification and description of each video frame into defects pixels or not, by means of a novel CNN-based encoder-decoder architecture, and 2) spatio-temporal analysis to group and fine-tune the detections. For blotch detection, the learned features, extracted from an intermediate stage of the network, are used to evaluate the dissimilarity between the pre-selected regions in consecutive frames. For scratches detection, the morphology of scratches is used to eliminate false alarms. The experiments are performed on various video sequences suffering synthetic and real scratches and blotches. The results demonstrate the effectiveness of our approach and significant improvement against the most recent detectors.
Historical document images suffer from several kinds of degradation affecting their readability a... more Historical document images suffer from several kinds of degradation affecting their readability and increase their sizes when digitizing. Indeed, preserving the old document heritage requires an important storage infrastructures, while this space could be significantly reduced if one manages to safeguard these documents without their considerable amount of artifacts which increasing their entropy. We present in this work a new and more advantageous solution for the document image compression technology dedicated to National Archives and Libraries. The aim of this work is to propose a novel historical document compression scheme judiciously combined with a novel automatic enhancement scheme, both based on PDE-Analysis applied in an innovative way, to face this challenging issue without affecting the look and feel of the historical document image. Throughout experiment, the proposal is fairly evaluated on overall DIBCO datasets and some historical documents collected from the web by means of objective measures and perceptual judgment against the most used enhancement algorithms, the most robust compression standards such as JPEG-XR and the most used sequential compression-enhancement combinations techniques to proof the robustness of the proposed collaborative approach against classic sequential approaches.
2016 International Symposium on Signal, Image, Video and Communications (ISIVC), 2016
In this paper, we firstly study the possible usability of two state-of-the-art no-reference image... more In this paper, we firstly study the possible usability of two state-of-the-art no-reference image quality assessment (IQA) metrics (NIQE and BIQES) originally designed for natural image, in the context of medical ones. We also proposed a modified NIQE, called NIQE-K, inspired by some BIQES features and more adapted to medical images. This study, based on evaluating concurrently the three methods, first encompasses tests conducted on natural IQA database (LIVE-Release2 and CSIQ). The second experiment is conducted on an ultrasound image with noise distortions. The last experiment includes tests on Magnetic Resonance images with compression distortions analyzed with quality scores evaluated by radiologists.
2017 IEEE International Conference on Image Processing (ICIP), 2017
Ultrasonographic examination, either as visual inspection or quantitative analysis, is less effec... more Ultrasonographic examination, either as visual inspection or quantitative analysis, is less effective than other medical imaging systems due to speckle noise. The state-of-the-art speckle reduction methods often offers an effective speckle reduction but generally they suffer from oversmoothig, blurring effect and man-made/artificial appearance. In this paper, a new Multi-Output Filter based on a Multiplicative Multiresolution Decomposition (MOF-MMD) is proposed. This multiscale based method, particularly efficient in the case of multiplicative noise, enhances distinctively three outputs: edges, texture and the global image. The multi-output filter aims at offering an enhanced images according to the features desired by radiologists. The different structures, textures and edges are filtered according to the contour image obtained by morphological operators. Finally, we compare the MOF-MMD method with two state-of-the-art speckle reduction methods in terms of speckle reduction capacity and image quality improvement. The results show that the proposed method offers an effective speckle reduction with an improvement of the image quality without blurry and over-smoothing effect.
This paper addresses the problem of blind image restoration. The non blind iterative Lucy-Richard... more This paper addresses the problem of blind image restoration. The non blind iterative Lucy-Richardson (LR) algorithm is considered. The proposed architecture solution consists of making the LR algorithm totally blind and fully automatic. For this aim, first a proposed point spread function estimation method is used to estimate the blur kernel and make the LR algorithm blind. Moreover in order to make the LR algorithm automatic, the iterative deconvolution process is controlled by a proposed stopping criterion based on a no reference blurred image quality metric that suggested in [1]. Computer simulations using the Gblur LIVE database [2] are provided in order to illustrate the proposed architecture effectiveness. A comparative study with the shock filter blind deblurring method [3] is performed, emphasizing the good proposed architecture solution behavior. Keywords-blur; deblurring; multiplicative multiresolution decomposition (MMD); inverse problem; blur quality assessment.
In this paper, we want to launch a first step for the classification of images according to their... more In this paper, we want to launch a first step for the classification of images according to their degree of blur based on the subjective measurement of DMOS image quality in the Gblur database. For this purpose, we have carried out a comparative study on several classifiers in order to build a robust learning model based on the transformation into a DCT. The class imbalanced has forced us to look for and find appropriate performance evaluation metrics so that the comparison is not biased. It has been found that random forests (RF) give the best overall performance but that other classifiers discriminate better between certain types of images (depending on the degree of blur) than others. Finally, we compared the classification by the proposed model with another classification based on the NIQE quality measurement algorithm. The results of the model proposed by its simplicity are very promising.
Journal of Visual Communication and Image Representation, Feb 1, 2019
In this work, we present a fully connected convolutional encoder-decoder for defects detection in... more In this work, we present a fully connected convolutional encoder-decoder for defects detection in archived video. The proposed method handles the detection of two of the most common archived video-related defects, namely blotches and scratches. It consists of two stages: 1) pixel-level classification and description of each video frame into defects pixels or not, by means of a novel CNN-based encoder-decoder architecture, and 2) spatio-temporal analysis to group and fine-tune the detections. For blotch detection, the learned features, extracted from an intermediate stage of the network, are used to evaluate the dissimilarity between the pre-selected regions in consecutive frames. For scratches detection, the morphology of scratches is used to eliminate false alarms. The experiments are performed on various video sequences suffering synthetic and real scratches and blotches. The results demonstrate the effectiveness of our approach and significant improvement against the most recent detectors.
Abstract. We propose a method for blotch detection in archived videos by modeling their spatiotem... more Abstract. We propose a method for blotch detection in archived videos by modeling their spatiotemporal properties. We introduce an adaptive spatiotemporal segmentation to extract candidate regions that can be classified as blotches. Then, the similarity between the preselected regions and their corresponding motion-compensated regions in the adjacent frames is assessed by means of motion trajectory estimation and textural information analysis. Perceived ground truth based on just noticeable contrast is employed for the evaluation of our approach against the state-of-the-art, and the reported results show a better performance for our approach.
2017 European Conference on Electrical Engineering and Computer Science (EECS)
Image segmentation is often required as a fundamental stage in microstructure material characteri... more Image segmentation is often required as a fundamental stage in microstructure material characterization. The objective of this work is to choose hybridization between the Level Set method and the clustering approach in order to extract the characteristics of the materials from the segmentation result of the micrographic images. More specifically, the proposed approach contains two successive necessary stages. The first one consists in the application of possibilistic c-means clustering approach (PCM) to get the various classes of the original image. The second stage is based on using the result of the clustering approach i.e. one class among the three existing classes (which interests us) as an initial contour of the level set method to extract the boundaries of interest region. The main purpose of using the result of the PCM algorithm as initial step of the level set method is to enhance and facilitate the work of the latter. Our experimental results on real micrographic images show that the proposed segmentation method can extract successfully the interest region according to the chosen class and confirm its efficiency for segmenting micrographic images of materials.
This paper turns on blotches detection in archived video. Blotches could be considered as the mos... more This paper turns on blotches detection in archived video. Blotches could be considered as the most common artifacts in archived video. This type of artifacts is not temporally correlated, it can be considered as impulse noise in time. To detect blotches, the proposed approach is performed into two mains steps: blotch regions candidate extraction and classification of these regions as blotches or not. The first step is performed by using sudden changes region detection by means of median filtering in time. The second step relies on finding similarity of the candidate region in the adjacent frames in gradient space. The proposed method has been compared against ROD, SROD and SDIx detectors. The obtained results show that our proposed approach outperforms the considered methods.
2017 International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)
Verification of individual identity through the process of biometric identification involves comp... more Verification of individual identity through the process of biometric identification involves comparison between an encoded value and a stored value of the biometric feature in question. The effectiveness of a multimodal user authentication system is greater, but so is its complexity. The system error rate is reduced by the fact that multiple biometric features are combined, thus solving the weakness of the single biometric. Performance of individual authentication through palm-print-and iris-based bimodal biometric system is proposed in the present study. To this end, Log-Gabor filter and BSIF (Binarised Statistical Image Feature) coefficients are employed to obtain the iris and palmprint traits, and subsequently selection of the features vector is conducted with mRMR (Minimum Redundancy Maximum Relevance) transforms in higher coefficients. To match the iris or palm-print feature vector, the Hamming Distance is applied. According to the experiment outcomes, the proposed system not only has a significantly high recognition rate but it also affords greater security compared to the single biometric system.
2017 International Conference on Advanced Systems and Electric Technologies (IC_ASET), 2017
In Automatic Fingerprint Identification System (AFIS), the performance highly depends on features... more In Automatic Fingerprint Identification System (AFIS), the performance highly depends on features extraction. In this paper, a new minutiae detection method based on wave atoms transform is presented. The proposed approach is based on analyzing fingerprints by specific representation highlighting the oscillatory structure. Thereby, a fine analysis from local characteristics via the block wave atoms transform is introduced. We explain how minutiae are detected by the use of wave atoms transform associated to a support vector machine (SVM) classifier. In order to evaluate the algorithm, FVC 2002 fingerprint databases have been considered. The tests show that for fingerprints features extraction, the wave atoms transform offers a good performance.
2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP), 2020
This paper deals with the detection of blurred images that may eventually be captured by a drone.... more This paper deals with the detection of blurred images that may eventually be captured by a drone. The embedded system should be able to measure the amount of blur affecting the images in order to decide whether to acquire the scene again or not. For this purpose, we have developed a simple model based on Discrete Cosine Transform (DCT) associated to Support Vector Machine Classifier SVM, to classify images into three categories and thus detect strongly, moderately and slightly blurred images. The proposed system has been tested on 550 images captured by a drone. The obtained results are very conclusive.
Palmprint recognition systems are dependent on feature extraction. A method of feature extraction... more Palmprint recognition systems are dependent on feature extraction. A method of feature extraction using higher discrimination information was developed to characterize palmprint images. In this method, two individual feature extraction techniques are applied to a discrete wavelet transform of a palmprint image, and their outputs are fused. The two techniques used in the fusion are the histogram of gradient and the binarized statistical image features. They are then evaluated using an extreme learning machine classifier before selecting a feature based on principal component analysis. Three palmprint databases,
Journal of Visual Communication and Image Representation, 2019
In this work, we present a fully connected convolutional encoder-decoder for defects detection in... more In this work, we present a fully connected convolutional encoder-decoder for defects detection in archived video. The proposed method handles the detection of two of the most common archived video-related defects, namely blotches and scratches. It consists of two stages: 1) pixel-level classification and description of each video frame into defects pixels or not, by means of a novel CNN-based encoder-decoder architecture, and 2) spatio-temporal analysis to group and fine-tune the detections. For blotch detection, the learned features, extracted from an intermediate stage of the network, are used to evaluate the dissimilarity between the pre-selected regions in consecutive frames. For scratches detection, the morphology of scratches is used to eliminate false alarms. The experiments are performed on various video sequences suffering synthetic and real scratches and blotches. The results demonstrate the effectiveness of our approach and significant improvement against the most recent detectors.
Historical document images suffer from several kinds of degradation affecting their readability a... more Historical document images suffer from several kinds of degradation affecting their readability and increase their sizes when digitizing. Indeed, preserving the old document heritage requires an important storage infrastructures, while this space could be significantly reduced if one manages to safeguard these documents without their considerable amount of artifacts which increasing their entropy. We present in this work a new and more advantageous solution for the document image compression technology dedicated to National Archives and Libraries. The aim of this work is to propose a novel historical document compression scheme judiciously combined with a novel automatic enhancement scheme, both based on PDE-Analysis applied in an innovative way, to face this challenging issue without affecting the look and feel of the historical document image. Throughout experiment, the proposal is fairly evaluated on overall DIBCO datasets and some historical documents collected from the web by means of objective measures and perceptual judgment against the most used enhancement algorithms, the most robust compression standards such as JPEG-XR and the most used sequential compression-enhancement combinations techniques to proof the robustness of the proposed collaborative approach against classic sequential approaches.
2016 International Symposium on Signal, Image, Video and Communications (ISIVC), 2016
In this paper, we firstly study the possible usability of two state-of-the-art no-reference image... more In this paper, we firstly study the possible usability of two state-of-the-art no-reference image quality assessment (IQA) metrics (NIQE and BIQES) originally designed for natural image, in the context of medical ones. We also proposed a modified NIQE, called NIQE-K, inspired by some BIQES features and more adapted to medical images. This study, based on evaluating concurrently the three methods, first encompasses tests conducted on natural IQA database (LIVE-Release2 and CSIQ). The second experiment is conducted on an ultrasound image with noise distortions. The last experiment includes tests on Magnetic Resonance images with compression distortions analyzed with quality scores evaluated by radiologists.
2017 IEEE International Conference on Image Processing (ICIP), 2017
Ultrasonographic examination, either as visual inspection or quantitative analysis, is less effec... more Ultrasonographic examination, either as visual inspection or quantitative analysis, is less effective than other medical imaging systems due to speckle noise. The state-of-the-art speckle reduction methods often offers an effective speckle reduction but generally they suffer from oversmoothig, blurring effect and man-made/artificial appearance. In this paper, a new Multi-Output Filter based on a Multiplicative Multiresolution Decomposition (MOF-MMD) is proposed. This multiscale based method, particularly efficient in the case of multiplicative noise, enhances distinctively three outputs: edges, texture and the global image. The multi-output filter aims at offering an enhanced images according to the features desired by radiologists. The different structures, textures and edges are filtered according to the contour image obtained by morphological operators. Finally, we compare the MOF-MMD method with two state-of-the-art speckle reduction methods in terms of speckle reduction capacity and image quality improvement. The results show that the proposed method offers an effective speckle reduction with an improvement of the image quality without blurry and over-smoothing effect.
This paper addresses the problem of blind image restoration. The non blind iterative Lucy-Richard... more This paper addresses the problem of blind image restoration. The non blind iterative Lucy-Richardson (LR) algorithm is considered. The proposed architecture solution consists of making the LR algorithm totally blind and fully automatic. For this aim, first a proposed point spread function estimation method is used to estimate the blur kernel and make the LR algorithm blind. Moreover in order to make the LR algorithm automatic, the iterative deconvolution process is controlled by a proposed stopping criterion based on a no reference blurred image quality metric that suggested in [1]. Computer simulations using the Gblur LIVE database [2] are provided in order to illustrate the proposed architecture effectiveness. A comparative study with the shock filter blind deblurring method [3] is performed, emphasizing the good proposed architecture solution behavior. Keywords-blur; deblurring; multiplicative multiresolution decomposition (MMD); inverse problem; blur quality assessment.
In this paper, we want to launch a first step for the classification of images according to their... more In this paper, we want to launch a first step for the classification of images according to their degree of blur based on the subjective measurement of DMOS image quality in the Gblur database. For this purpose, we have carried out a comparative study on several classifiers in order to build a robust learning model based on the transformation into a DCT. The class imbalanced has forced us to look for and find appropriate performance evaluation metrics so that the comparison is not biased. It has been found that random forests (RF) give the best overall performance but that other classifiers discriminate better between certain types of images (depending on the degree of blur) than others. Finally, we compared the classification by the proposed model with another classification based on the NIQE quality measurement algorithm. The results of the model proposed by its simplicity are very promising.
Uploads
Papers by amina Serir