Yaakoub Boualleg

Diabetic retinopathy (DR) is a leading cause of blindness worldwide, necessitating early detection to prevent severe visual impairment. Despite numerous proposed classification techniques, challenges persist due to the high parameter count of deep learning algorithms, imbalanced datasets, and limited performance. This study introduces a novel framework for DR classification that leverages multi-view deep features, multilinear whitened principal component analysis, tensor exponential discriminant analysis, synthetic minority oversampling technique, and deep random forest. We evaluated this architecture using the APTOS blindness dataset under a standard protocol. The results demonstrate that our architecture significantly improves classification accuracy, surpassing existing methods. Our contributions highlight a promising approach for enhancing DR classification performance.

Comparing Pre-trained Language Model for Arabic Hate Speech Detection

Nowadays, hate speech detection from Arabic tweets attracts the attention of many researchers. Numerous systems and techniques have been proposed to address this classification challenge. Nonetheless, three major limits persist: the use of deep learning models with an excess of hyperparameters, the reliance on hand-crafted features, and the requirement for a huge amount of training data to achieve satisfactory performance. In this study, we propose Contextual Deep Random Forest (CDRF), a hate speech detection approach that combines contextual embedding and Deep Random Forest. From the experimental findings, the Arabic contextual embedding model proves to be highly effective in hate speech detection, outperforming the static embedding models. Additionally, we prove that the proposed CDRF significantly enhances the performance of Arabic hate speech classification.

Remote Sensing Image Retrieval (RSIR) is quite a challenging topic due to the complexity of High-Resolution Remote Sensing (HRRS) imagery. Recently, most researches are focusing on feature extraction using learned features instead of traditional hand-crafted features for RSIR. However, on the one hand, learning discriminative feature representations is not sufficient due to the limited ability of the standard metric distances for matching similar images. On the other hand, the high dimensionality of the learned features makes the retrieval process not efficient for large-scale collections of HRRS images. To address these issues, we propose a Triplet Low-Dimensional Convolutional Neural Network (TLDCNN) which is based on the LDCNN model combined with triplet loss function. We use end-to-end training process to automatically learn fine-grained image similarity directly from input triplet images. The experiment results demonstrated the effectiveness of the proposed network on three publicly available HRRS image datasets for HRRSIR.

High-resolution remote sensing scene classification (HR-RSSC) plays an increasingly important role since it aims to enhance the scene semantic understanding. Recently, convolutional neural networks (CNNs) proved their effectiveness in learning powerful feature representations for various visual recognition tasks. However, in the RS domain, the performance of CNN is still limited due to the lack of sufficient labeled data. In this letter, we propose an HR-RSSC method based on CNN transfer learning (TL) for feature extraction (FE) and deep forest (DF) for classification. In fact, we extract deep features from the last convolutional layer in order to avoid the use of the fully connected layers (FCLs) which need many parameters to tune. Moreover, we train a DF model that is based on ensemble learning that can achieve better performances than single classifiers and is easy to train with few parameters. We evaluate the proposed method on two RS image data sets. Compared to full-training, fine-tuning, and state-of-the-art CNN TL methods, the results demonstrate the effectiveness of the DF model for HR-RSSC based on CNN TL in terms of overall accuracy and training time.

In this paper, we address the semantic gap problem in high-spatial resolution remote sensing images retrieval. We propose a useful semantic image representation that improves the understanding of the machine with respect to the human perception. We use a remote sensing scene classification Convolutional Neural Network (CNN) model to detect the semantic concepts. The similarity distance is calculated to retrieve the most similar images to the given query image. Then, to improve the performance of the retrieval results, a relevance feedback phase has been proposed, which ensures that the final result corresponds to the user need. Our proposal shows promising results and improves the retrieved quality with respect to state-of-the-art approaches.

In this work, we contribute to solve the community detection problem by proposing an algorithm for the detection of disjoint communities' cores considering unweighted and undirected social graphs. The proposed algorithm is based on the removal of non-essential edges and induced isolated nodes from networks (graphs). To this purpose; we have built a model for predicting edges removal using Weighted Support Vector Machines (WSVM) trained on real-life social network datasets. The training phase was carried out by means of an appropriate proposed heuristic to label edges, and discriminating features extracted from both edges end-point nodes and network structure characteristics. Our designed algorithm shows promising results for communities' cores detection.