Exploring Complementarity and Explainability in CNNs for Periocular Verification Across Acquisition Distances

A recent study delves into the effectiveness of various convolutional neural networks (CNNs) for periocular verification, particularly focusing on how distance affects performance. By training models like SqueezeNet, MobileNetv2, and ResNet50 on eye crops from the VGGFace2 dataset, researchers are pushing the boundaries of facial recognition technology. This research is significant as it not only enhances the accuracy of biometric systems but also explores innovative methods like LIME heatmaps for better interpretability, making AI more transparent and reliable.