Abstract
Recognition using ear images has been an active field of research in recent years. Besides faces and fingerprints, ears have a unique structure to identify people and can be captured from a distance, contactless, and without the subject’s cooperation. Therefore, it represents an appealing choice for building surveillance, forensic, and security applications. However, many techniques used in those applications—e.g., convolutional neural networks (CNN)—usually demand large-scale datasets for training. This research work introduces a new dataset of ear images taken under uncontrolled conditions that present high inter-class and intra-class variability. We built this dataset using an existing face dataset called the VGGFace, which gathers more than 3.3 million images. in addition, we perform ear recognition using transfer learning with CNN pretrained on image and face recognition. Finally, we performed two experiments on two unconstrained datasets and reported our results using Rank-based metrics.
Highlights
Identifying people is a persistent issue in society
Among the most common physical biometric traits such as fingerprints, palmprints, hand geometry, iris, and face, the ear structure results in an excellent source to identify a person without their cooperation
Unlike the University of Beira Interior Ear Dataset (UBEAR) dataset, the Annotated Web Ears (AWE) for Segmentation dataset [17,18] collects images taken under unconstrained environments since they were gathered from the web and of different resolutions
Summary
Identifying people is a persistent issue in society. Different areas such as forensic science, surveillance, and security systems usually demand solutions for this issue. Among the most common physical biometric traits such as fingerprints, palmprints, hand geometry, iris, and face, the ear structure results in an excellent source to identify a person without their cooperation. In our case, using one of the most extensive face datasets available, the VGGFace2 [12], we built and proposed an extended ear dataset for training and testing approaches in the unconstrained ear recognition context. We fine-tune pretrained models on general and face images to adapt the models to the ear recognition domain using the proposed dataset. In a previous version of this paper [15], we fine-tune models using a reduced part of the VGGFace dataset In this version, we present the completed extended version of the VGGFace-Ear dataset, which was used to train and test models according to the proposal.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have