Multimodal Approach for Enhancing Biometric Authentication.

Unimodal biometric systems rely on a single source or unique individual biological trait for measurement and examination. Fingerprint-based biometric systems are the most common, but they are vulnerable to presentation attacks or spoofing when a fake fingerprint is presented to the sensor. To address this issue, we propose an enhanced biometric system based on a multimodal approach using two types of biological traits. We propose to combine fingerprint and Electrocardiogram (ECG) signals to mitigate spoofing attacks. Specifically, we design a multimodal deep learning architecture that accepts fingerprints and ECG as inputs and fuses the feature vectors using stacking and channel-wise approaches. The feature extraction backbone of the architecture is based on data-efficient transformers. The experimental results demonstrate the promising capabilities of the proposed approach in enhancing the robustness of the system to presentation attacks.

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.

COVID-19 Detection in CT/X-ray Imagery Using Vision Transformers.

The steady spread of the 2019 Coronavirus disease has brought about human and economic losses, imposing a new lifestyle across the world. On this point, medical imaging tests such as computed tomography (CT) and X-ray have demonstrated a sound screening potential. Deep learning methodologies have evidenced superior image analysis capabilities with respect to prior handcrafted counterparts. In this paper, we propose a novel deep learning framework for Coronavirus detection using CT and X-ray images. In particular, a Vision Transformer architecture is adopted as a backbone in the proposed network, in which a Siamese encoder is utilized. The latter is composed of two branches: one for processing the original image and another for processing an augmented view of the original image. The input images are divided into patches and fed through the encoder. The proposed framework is evaluated on public CT and X-ray datasets. The proposed system confirms its superiority over state-of-the-art methods on CT and X-ray data in terms of accuracy, precision, recall, specificity, and F1 score. Furthermore, the proposed system also exhibits good robustness when a small portion of training data is allocated.

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.

Transformers and Generative Adversarial Networks for Liveness Detection in Multitarget Fingerprint Sensors.

Fingerprint-based biometric systems have grown rapidly as they are used for various applications including mobile payments, international border security, and financial transactions. The widespread nature of these systems renders them vulnerable to presentation attacks. Hence, improving the generalization ability of fingerprint presentation attack detection (PAD) in cross-sensor and cross-material setting is of primary importance. In this work, we propose a solution based on a transformers and generative adversarial networks (GANs). Our aim is to reduce the distribution shift between fingerprint representations coming from multiple target sensors. In the experiments, we validate the proposed methodology on the public LivDet2015 dataset provided by the liveness detection competition. The experimental results show that the proposed architecture yields an increase in average classification accuracy from 68.52% up to 83.12% after adaptation.

Rhythm-based heartbeat duration normalization for atrial fibrillation detection.

BACKGROUND: Screening of atrial fibrillation (AF) for high-risk patients including all patients aged 65 years and older is important for prevention of risk of stroke. Different technologies such as modified blood pressure monitor, single lead ECG-based finger-probe, and smart phone using plethysmogram signal have been emerging for this purpose. All these technologies use irregularity of heartbeat duration as a feature for AF detection. We have investigated a normalization method of heartbeat duration for improved AF detection. METHOD: AF is an arrhythmia in which heartbeat duration generally becomes irregularly irregular. From a window of heartbeat duration, we estimate the possible rhythm of the majority of heartbeats and normalize duration of all heartbeats in the window based on the rhythm so that we can measure the irregularity of heartbeats for both AF and non-AF rhythms in the same scale. Irregularity is measured by the entropy of distribution of the normalized duration. Then we classify a window of heartbeats as AF or non-AF by thresholding the measured irregularity. The effect of this normalization is evaluated by comparing AF detection performances using duration with the normalization, without normalization, and with other existing normalizations. RESULTS: Sensitivity and specificity of AF detection using normalized heartbeat duration were tested on two landmark databases available online and compared with results of other methods (with/without normalization) by receiver operating characteristic (ROC) curves. ROC analysis showed that the normalization was able to improve the performance of AF detection and it was consistent for a wide range of sensitivity and specificity for use of different thresholds. Detection accuracy was also computed for equal rates of sensitivity and specificity for different methods. Using normalized heartbeat duration, we obtained 96.38% accuracy which is more than 4% improvement compared to AF detection without normalization. CONCLUSIONS: The proposed normalization method was found useful for improving performance and robustness of AF detection. Incorporation of this method in a screening device could be crucial to reduce the risk of AF-related stroke. In general, the incorporation of the rhythm-based normalization in an AF detection method seems important for developing a robust AF screening device.

Evaluating Belief Structure Satisfaction to Uncertain Target Values.

We describe the basic properties of the Dempster-Shafer belief structure and introduce the associated measures of plausibility and belief. We look at the role of these structures for providing a model of imprecise probabilistic information. We next consider the problem of calculating the satisfaction of target values by a variable V whose value is expressed by a belief structure. We first look at the simplest case when the target is expressed as subset of the domain of V . We then look at the situation when the target is expressed by more complex uncertain structures. Among those considered are a probability distribution, another belief structure, measure, and possibility distribution. At a formal level this paper involves the extension of the concepts of plausibility and belief associated with D-S structures from being mappings of subsets of the underlying domain of V into unit interval to be mappings of these more complex structures into the unit interval.

HBS: a novel biometric feature based on heartbeat morphology.

In this paper, a new feature named heartbeat shape (HBS) is proposed for ECG-based biometrics. HBS is computed from the morphology of segmented heartbeats. Computation of the feature involves three basic steps: 1) resampling and normalization of a heartbeat; 2) reduction of matching error; and 3) shift invariant transformation. In order to construct both gallery and probe templates, a few consecutive heartbeats which could be captured in a reasonably short period of time are required. Thus, the identification and verification methods become efficient. We have tested the proposed feature independently on two publicly available databases with 76 and 26 subjects, respectively, for identification and verification. The second database contains several subjects having clinically proven cardiac irregularities (atrial premature contraction arrhythmia). Experiments on these two databases yielded high identification accuracy (98% and 99.85%, respectively) and low verification equal error rate (1.88% and 0.38%, respectively). These results were obtained by using templates constructed from five consecutive heartbeats only. This feature compresses the original ECG signal significantly to be useful for efficient communication and access of information in telecardiology scenarios.

Geometry-based image retrieval in binary image databases.

In this paper, a geometry-based image retrieval system is developed for multi-object images. We model both shape and topology of image objects using a structured representation called curvature tree (CT). The hierarchy of the CT reflects the inclusion relationships between the image objects. To facilitate shape-based matching, triangle-area representation (TAR) of each object is stored at the corresponding node in the CT. The similarity between two multi-object images is measured based on the maximum similarity subtree isomorphism (MSSI) between their CTs. For this purpose, we adopt a recursive algorithm to solve the MSSI problem and a very effective dynamic programming algorithm to measure the similarity between the attributed nodes. Our matching scheme agrees with many recent findings in psychology about the human perception of multi-object images. Experiments on a database of 13500 real and synthesized medical images and the MPEG-7 CE-1 database of 1400 shape images have shown the effectiveness of the proposed method.