Biometric technologies are fast becoming a requirement in security systems today, providing solutions where traditional means alone would not be adequate. This paper proposes FootprintNet, a Siamese network that utilizes pre-trained convolutional neural networks, specifically EfficientNet, MobileNet, and ShuffleNet, to improve the robustness and accuracy of footprint recognition. By learning the ability to identify fine distinctions between images of footprints, FootprintNet offers great biometric identification potential. Detailed analysis of the Biometric 220 × 6 Human Footprint dataset shows a true positive rate over 99% under various thresholds and a precision rate of 100% during training. Most importantly, this system is also applicable to newborn and infant identification, making it especially significant in medical settings, including hospitals and birthing clinics. Furthermore, the model sizes range from 7.8 MB (ShuffleNet) to 24.5 MB (EfficientNet), which makes FootprintNet deployable on low-computational-power devices—a highly desirable trait for mobile or high-security applications.

Today, automatic identification of individuals from biometric features is widely used in identification and authentication, security, and monitoring applications. Since facial recognition is a more user-friendly and comfortable method than other biometric methods, it has grown rapidly in recent years. However, most facial recognition systems are vulnerable to spoofing attacks. Therefore, face liveness detection (FLD) methods are of great importance. On the other hand, unlike traditional methods, deep learning techniques promise to significantly increase the accuracy of facial liveness detection systems and eliminate the difficulties of the real-world implementation of these systems. Therefore, in this paper, the application of some deep learning models to detect face liveness is reviewed and compared with each other.

Palmprint recognition (PR) has garnered significant interest due to its distinctive characteristics and potential applications as a robust and secure biometric authentication (BA) method across various domains. This paper introduces a novel approach to enhance PR using pre-trained Convolutional Neural Network (CNN) models. The proposed method process harnesses the capabilities of Deep Learning (DL) and Transfer Learning (TL) to improve palmprint recognition by leveraging state-of-the-art CNN architectures for feature extraction and classification. This proposed method begins by exploring the potential of pre-trained CNN architectures, including ShuffleNet, EfficientNet, and MobileNet, as feature extractors for palmprints. In the next step, extracted feature vectors are compared using the Cosine Similarity (CS) method. The proposed method is thoroughly evaluated through comprehensive experiments. Results of evaluations demonstrate that pre-trained CNN models excel at recognizing palmprints for biometric authentication, establishing their proficiency in this domain. Consequently, this paper illuminates the inherent capabilities of pre-trained CNN models as a potent tool for advancing PR, introducing an innovative facet to BA methodologies.