Publications Scientifiques
Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10
Browse
20 results
Search Results
Item High-Quality Synthesized Face Sketch Using Generative Reference Prior(Polska Akademia Nauk, 2024) Mahfoud, Sami; Bengherabi, Messaoud; Daamouche, Abdelhamid; Boutellaa, Elhocine; Hadid, AbdenourFace sketch synthesis (FSS) is considered as an image-to-image translation problem, where a face sketch is generated from an input face photo. FSS plays a vital role in video/image surveillance-based law enforcement. In this paper, motivated by the recent success of generative adversarial networks (GAN), we consider conditional GAN (cGAN) to approach the problem of face sketch synthesis. However, despite the powerful cGAN model’s ability to generate fine textures, low-quality inputs characterized by the facial sketches drawn by artists cannot offer realistic and faithful details and have unknown degradation due to the drawing process, while high-quality references are inacces- sible or even unexistent. In this context, we propose an approach based on Generative Reference Prior (GRP) to improve the synthesized face sketch perception. Our proposed model, that we call cGAN-GRP, leverages diverse and rich priors encapsulated in a pre-trained face GAN for generating high-quality facial sketch synthesis. Extensive experiments on publicly available face databases using facial sketch recognition rate and image quality assessment metrics as criteria demonstrate the effectiveness of our proposed model compared to several state-of-the-art methods.Item Convolutional Encoder-Decoder Network for Road Extraction from Remote Sensing Images(Institute of Electrical and Electronics Engineers, 2024) Makhlouf, Yasmine; Daamouche, Abdelhamid; Melgani, FaridIn this paper, we propose a convolutional neural network, which is based on down sampling followed by up sampling architecture for the purpose of road extraction from aerial images. Our model consists of convolutional layers only. The proposed encoder-decoder structure allows our network to retain boundary information, which is a critical feature for road identification. This feature is usually lost when dealing with other CNN models. Our design is also less complex in terms of depth, number of parameters, and memory size. It, therefore, uses fewer computer resources in both training and during execution. Experimental results on Massachusetts roads dataset demonstrate that the proposed architecture, although less complex, competes with the state-of-the-art proposed approaches in terms of precision, recall, and accuracy.Item SIFT and Gabor Based Features Extraction Method Applied to the Infrared VHR Image of Boumerdes(Institute of Electrical and Electronics Engineers, 2024) Fiala, Chahrazed; Daamouche, AbdelhamidThis paper investigates the contribution of SIFT and Gabor features in the classification of a VHR Image, focusing on the Infrared and RGB channels. The assessment of the classification scheme is conducted by the SVM classifier, and the comparison using the average accuracy as a metric reveals that the infrared channel improves the performance of the classification scheme.Item CNN and M-SLIC superpixels feature fusion for VHR image classification(IEEE, 2022) Semcheddine, Belkis Asma; Daamouche, AbdelhamidIn this letter, we present a method for fusing handcrafted features with abstract features for the purpose of VHR remote sensing image classification. The proposed strategy allows for a multi-level feature fusion, which enriches the available spectral data, resulting in a better class separability. In a first step, deep features are extracted using Convolutional Neural Networks. These features are then fused with Haralick features drawn out by means of M-SLIC superpixels segmentation. The combined features are then concatenated with the spectral features of the image and classified using Support Vector Machines. Our experiments were conducted on a VHR satellite image, and the obtained results qualify us to validate the superiority of the suggested scheme (over 16% overall classification accuracy improvement)Item Optimization of matched filters for VHR image segmentation(Taylor and Francis, 2022) Semcheddine, Belkis Asma; Daamouche, AbdelhamidThe availability of very high resolution remotely sensed images has made the use of spectral information alone insufficient for class recognition, so the integration of spatial information into the segmentation task becomes necessary. Spatial information takes into account the surrounding of a pixel, rather than dealing with a pixel as an isolated item. In this paper, inspired by matched filtering theory, we propose a new technique for spatial feature extraction. The technique consists of using 2D kernels convolved with the spectral bands. The convolution operation bears valuable spatial information about the pixel under consideration. The concatenation of the extracted spatial features and spectral features of pixels feeds a support vector machine (SVM) to segment the image of interest. To cope with the complexity of images having objects of varying sizes (i.e. urban areas), we adopted a hierarchical strategy. That is, kernels with increasing size are applied to extract different spatial features corresponding to different objects. Thus, we have extended the use of matched filters from single object enhancement to multi-object enhancement. A challenging step in designing the matched filters is the two-dimensional kernels coefficients selection, which we formulated as an optimization problem within a particle swarm optimization framework. The optimization process was driven by two different fitness functions, the cross-validation SVM accuracy and the Bhattacharyya distance, which were both evaluated on training samples. We assessed the proposed procedure on two very high-resolution images having different spatial resolutions. The obtained results showed significant improvement in terms of the overall classification accuracy (over 10% for both images). Moreover, visual inspection of the segmented images, in addition to pepper and salt elimination, revealed significant improvement in many objects not detected by the spectral method. Our method of extracting spatial features seems to be very efficientItem A deep neural network approach to QRS detection using autoencoders(Elsevier, 2021) Belkadi, Mohamed Amine; Daamouche, Abdelhamid; Melgani, FaridObjective: In this paper, a stacked autoencoder deep neural network is proposed to extract the QRS complex from raw ECG signals without any conventional feature extraction phase. Methods: A simple architecture has been deeply trained on many datasets to ensure the generalization of the network at inference. Results: The proposed method achieved a QRS detection accuracy of 99.6% using more than 1042000 beats which is competitive with all state-of-the-art QRS detectors. Moreover, the proposed method produced only 0.82% of Detection Error Rate using six unseen datasets containing more than 1470000 beats. Thus confirms the high performance of our method to detect QRSs. Conclusion: Stacked autoencoder neural networks are very effective in QRS detection. At inference, our algorithm processes 1042309 beats in less than 25.32 s. Thus, it is favorably comparable with state-of-the-art deep learning methods. Significance: The stacked autoencoder is an efficient tool for QRS detection, which could replace conventional systems to help practitioners make fast and accurate decisionsItem A robust QRS detection approach using stationary wavelet transform(Springer, 2021) Belkadi, Mohamed Amine; Daamouche, AbdelhamidAccurate QRS detection is crucial for reliable ECG signal analysis and the development of automatic diagnosis tools. In this paper, we propose a simple yet efficient new algorithm for QRS detection using the Stationary Wavelet Transform (SWT). The wavelet transform has been extensively exploited for QRS detection and proved to be an efficient mathematical tool for scale analysis; it provides good frequency components estimation for the input signal and has good localization capability. The proposed procedure exploits solely the first level approximation coefficients of the wavelet transform applied to the bandpass-filtered ECG signal. Therefore, it resulted in a reduced complexity algorithm compared to the existing methods which use many decomposition levels. Thresholding has been implemented using the Pan-Tompkins procedure which is known to be very powerful. Our approach has been assessed over the MIT/BIH benchmark database, the MIT noise stress test database for noise robustness evaluation and the European ST-T database. The obtained results show competitive performance with state-of-the-art algorithms. The proposed scheme achieved a sensitivity of 99.83%, a positive predictivity of 99.94% and a detection error rate of 0.228% using Lead I MIT-BIH Database, this performance is one of the best results over this benchmark, and 99.35% of sensitivity, 99.76% of positive predictivity and detection error rate of 0.9% using the European ST-T Database, hence, our algorithm achieved high performance on Holter environment. Using the MIT noise stress test database, our algorithm achieved 98.77% of sensitivity, 91.01% of positive predictivity, and 10.12% of DER. Thus, our algorithm is robust and outperforms state-of-the-art algorithms on noisy recordingsItem IEEE GRSS Algeria chapter activities and perspectives [Chapters](IEEE, 2019) Karoui, Moussa Sofiane; Souissi, Boularbah; Daamouche, AbdelhamidThe IEEE Geoscience and Remote Sensing Society (GRSS) Algeria Chapter is the first such Chapter in the Maghreb and North Africa regions, the second on the African continent after the South Africa Chapter, and the third in the Arab world after those of Saudi Arabia and the United Arab Emirates. This Chapter was created through the joint participation of Algerian Space Agency (ASAL) researchers [in particular, researchers from the Center des Techniques Spatiales (CTS), an operational entity of the ASAL] and Algerian scientists participating in GRS fields from various Algerian universities, such as the Université des Sciences et de la Technologie Houari Boumedienne (USTHB), the Université M'Hamed Bougara (UMB), and the Université Djilali Liabès (UDL) (Figure 1). The research interests of this community range from the design of Earth observation missions (Algerian AlSat satellite missions) through the processing of remote sensing and geospatial data and the exploitation of those data in socioeconomic applicationsItem A new method for accurate QRS detection using stationary wavelet transform(Mohamed Amine Belkadi;, 2017) Belkadi, Mohamed Amine; Daamouche, AbdelhamidIt is well-known that the wavelet transform is a very useful mathematical tool for scale analysis, with very accurate frequency components estimation for the input signal. In this paper, we propose a new efficient method for QRS detection by employing the Stationary Wavelet Transform (SWT) also known as short wavelet transform. Our approach has been tested over MIT/BIH benchmark database. The obtained results are in a good agreement with the published works. Globally, we achieved a sensitivity of 99.733%, specificity of 99.922% and an error rate of 0.345% using Lead I ECGItem ECG as a biometric for individual's identification(IEEE, 2017) Sellami, Abdelkader; Zouaghi, Amine; Daamouche, AbdelhamidIn this paper, we investigate a new method to analyze electrocardiogram (ECG) signal, extract the features, for the real time human identification using single lead human electrocardiogram. The proposed system extracts special parts of the ECG signal starting from the P wave, the QRS complex and ending with the T wave for that we used the multiresolution wavelet analysis. Different features are selected and reconstructed from both amplitude and time interval of the ECG signal. The matching decisions are evaluated on the basis of correlation coefficient between the features and the Radial Basis function network classifier is introduced for validation and comparison. The performance evaluation was carried out on four ECG public databases with a total of 149 persons subjected to different physical activities and heart conditions, the preliminary results indicate that the system achieved an accuracy of 90-93%