Self-supervised learning for remote sensing scene classification under the few shot scenario.

Scene classification is a crucial research problem in remote sensing (RS) that has attracted many researchers recently. It has many challenges due to multiple issues, such as: the complexity of remote sensing scenes, the classes overlapping (as a scene may contain objects that belong to foreign classes), and the difficulty of gaining sufficient labeled scenes. Deep learning (DL) solutions and in particular convolutional neural networks (CNN) are now state-of-the-art solution in RS scene classification; however, CNN models need huge amounts of annotated data, which can be costly and time-consuming. On the other hand, it is relatively easy to acquire large amounts of unlabeled images. Recently, Self-Supervised Learning (SSL) is proposed as a method that can learn from unlabeled images, potentially reducing the need for labeling. In this work, we propose a deep SSL method, called RS-FewShotSSL, for RS scene classification under the few shot scenario when we only have a few (less than 20) labeled scenes per class. Under this scenario, typical DL solutions that fine-tune CNN models, pre-trained on the ImageNet dataset, fail dramatically. In the SSL paradigm, a DL model is pre-trained from scratch during the pretext task using the large amounts of unlabeled scenes. Then, during the main or the so-called downstream task, the model is fine-tuned on the labeled scenes. Our proposed RS-FewShotSSL solution is composed of an online network and a target network both using the EfficientNet-B3 CNN model as a feature encoder backbone. During the pretext task, RS-FewShotSSL learns discriminative features from the unlabeled images using cross-view contrastive learning. Different views are generated from each image using geometric transformations and passed to the online and target networks. Then, the whole model is optimized by minimizing the cross-view distance between the online and target networks. To address the problem of limited computation resources available to us, our proposed method uses a novel DL architecture that can be trained using both high-resolution and low-resolution images. During the pretext task, RS-FewShotSSL is trained using low-resolution images, thereby, allowing for larger batch sizes which significantly boosts the performance of the proposed pipeline on the task of RS classification. In the downstream task, the target network is discarded, and the online network is fine-tuned using the few labeled shots or scenes. Here, we use smaller batches of both high-resolution and low-resolution images. This architecture allows RS-FewshotSSL to benefit from both large batch sizes and full image sizes, thereby learning from the large amounts of unlabeled data in an effective way. We tested RS-FewShotSSL on three RS public datasets, and it demonstrated a significant improvement compared to other state-of-the-art methods such as: SimCLR, MoCo, BYOL and IDSSL.

LwF-ECG: Learning-without-forgetting approach for electrocardiogram heartbeat classification based on memory with task selector.

Most existing Electrocardiogram (ECG) classification methods assume that all arrhythmia classes are known during the training phase. In this paper, the problem of learning several successive tasks is addressed, where, in each new task, there are new arrhythmia classes to learn. Unfortunately, in machine learning it is known that when a model is retrained onto a new task, the machine tends to forget the old task. This is known in machine learning, as 'the catastrophic forgetting phenomenon'. To this end, a learn-without-forgetting (LwF) approach to solve this problem is proposed. This novel deep LwF method for ECG heartbeat classification is the first work of its kind in the field. This proposed LwF approach consists of a deep learning architecture that includes the following important aspects: feature extraction module, classification layers for each learned task, memory module to store one prototype for each task, and a task selection module able to identify the most suitable task for each input sample. The feature extraction module constitutes another contribution of this work. It starts with a set of deep layers that convert an ECG heartbeat signal into an image, then the pre-trained DenseNet169 CNN takes the obtained image and extracts rich and powerful features that are effective inputs for the classifications layers of the model. Whenever a new task is to be learned, the network expands with a new classification layer having a Softmax activation function. The newly added layer is responsible for learning the classes of the new task. When the network is trained for the new task, the shared layers, as well as the output layers of the old tasks, are also fine-tuned using pseudo labels. This helps in retaining knowledge of old tasks. Finally, the task selector stores feature prototypes for each task, and using a distance matching network, is trained to select which task is more suitable to classify a new test sample. The whole network uses end-to-end learning to optimize one loss functions, which is a weighted combination of the loss functions of the different network modules. The proposed model was tested on three common ECG datasets, namely the MIT-BIH, INCART, and SVDB datasets. The results obtained demonstrate the success of the proposed method in learning, without forgetting, successive ECG heartbeat classification tasks.

Unified Generative Adversarial Networks for Multidomain Fingerprint Presentation Attack Detection.

With the rapid growth of fingerprint-based biometric systems, it is essential to ensure the security and reliability of the deployed algorithms. Indeed, the security vulnerability of these systems has been widely recognized. Thus, it is critical to enhance the generalization ability of fingerprint presentation attack detection (PAD) cross-sensor and cross-material settings. In this work, we propose a novel solution for addressing the case of a single source domain (sensor) with large labeled real/fake fingerprint images and multiple target domains (sensors) with only few real images obtained from different sensors. Our aim is to build a model that leverages the limited sample issues in all target domains by transferring knowledge from the source domain. To this end, we train a unified generative adversarial network (UGAN) for multidomain conversion to learn several mappings between all domains. This allows us to generate additional synthetic images for the target domains from the source domain to reduce the distribution shift between fingerprint representations. Then, we train a scale compound network (EfficientNetV2) coupled with multiple head classifiers (one classifier for each domain) using the source domain and the translated images. The outputs of these classifiers are then aggregated using an additional fusion layer with learnable weights. In the experiments, we validate the proposed methodology on the public LivDet2015 dataset. The experimental results show that the proposed method improves the average classification accuracy over twelve classification scenarios from 67.80 to 80.44% after adaptation.

Electrocardiogram heartbeat classification based on a deep convolutional neural network and focal loss.

The electrocardiogram (ECG) is an effective tool for cardiovascular disease diagnosis and arrhythmia detection. Most methods proposed in the literature include the following steps: 1) denoizing, 2) segmentation into heartbeats, 3) feature extraction, and 4) classification. In this paper, we present a deep learning method based on a convolutional neural network (CNN) model. CNN models can perform feature extraction automatically and jointly with the classification step. In other words, our proposed method does not require a feature extraction step with hand-crafted techniques. Our proposed method is also based on an algorithm for heartbeat segmentation that is different from most existing methods. In particular, the segmentation algorithm defines each ECG heartbeat to start at an R-peak and end after 1.2 times the median RR time interval in a 10-s window. This method is simple and effective, as it does not use any form of filtering or processing that requires assumptions about the signal morphology or spectrum. Although enhanced ECG heartbeat classification algorithms have been proposed in the literature, they failed to achieve high performance in detecting some heartbeat categories, especially for imbalanced datasets. To overcome this challenge, we propose an optimization step for the deep CNN model using a novel loss function called focal loss. This function focuses on minority heartbeat classes by increasing their importance. We trained and evaluated our proposed model with the MIT-BIH and INCART datasets to identify five arrhythmia categories (N, S, V, Q, and F) based on the Association for Advancement of Medical Instrumentation (AAMI) standard. The evaluation results revealed that the focal loss function improved the classification accuracy for the minority classes as well as the overall metrics. Our proposed method achieved 98.41% overall accuracy, 98.38% overall F1-score, 98.37% overall precision, and 98.41% overall recall. In addition, our method achieved better performance than that of existing state-of-the-art methods.