The accuracy of face recognition systems is significantly affected by the quality of face sample images. There are many existing no-reference image quality metrics (IQMs) that are able to assess natural image quality by taking into account similar image-based quality attributes. Previous study showed that IQMs can assess face sample quality according to the biometric system performance. In addition, re-training an IQM can improve its performance for face biometric images. However, only one database was used in the previous study, and it contains only image-based distortions. In this paper, we propose to extend the previous study by use multiple face database including FERET color face database, and apply multiple setups for the re-training process in order to investigate how the re-training process affect the performance of no-reference image quality metric for face biometric images. The experimental results show that the performance of the appropriate IQM can be improved for multiple databases, and different re-training setups can influence the IQM’s performance.
The quality of a biometric sample is one of the main criteria having a direct influence on the overall performance of a biometric system. There are many existing researches focusing on biometric sample quality assessment, but different evaluation approaches measure different quality attributes and most of them focus on measuring modality-based attributes. Meanwhile, different biometric modalities seem to be isolated from each other in the image quality evaluation process. Quality metrics that can evaluate multi-modality biometric sample quality is rarely considered. The link of sample quality evaluation between different modalities can be established by using image-based quality metrics, which are able to assess image-based quality attributes. This could be the solution of developing multi-modality biometric sample quality evaluation approaches especially when the fingerprint acquisition sensor becomes contactless. In order to investigate the common framework of biometric sample quality assessment between contactless fingerprint, face, and iris, we will first review the commonly used image-based quality attributes for three modalities by surveying existing literature. Based on the survey, we identify and categorize these attributes to propose a refined selection of important ones for the assessment of multi-modality biometric sample quality.
Pathology detection in Wireless Capsule Endoscopy (WCE) using deep learning has been explored in the recent past. However, deep learning models can be influenced by the color quality of the dataset used to train them, impacting detection, segmentation and classification tasks. In this work, we evaluate the impact of color correction on pathology detection using two prominent object detection models: Retinanet and YOLOv5. We first generate two color corrected versions of a popular WCE dataset (i.e., SEE-AI dataset) using two different color correction functions. We then evaluate the performance of the Retinanet and YOLOv5 on the original and color corrected versions of the dataset. The results reveal that color correction makes the models generate larger bounding boxes and larger intersection areas with the ground truth annotations. Furthermore, color correction leads to an increased number of false positives for certain pathologies. However, these effects do not translate into a consistent improvement in performance metrics such as F1-scores, IoU, and AP50. The code is available at https://github.com/agossouema2011/WCE2024. Keywords: Wireless Capsule Endoscopy, Color correction, Retinanet, YOLOv5, Detection
Image quality assessment plays an important role in iris recognition systems because the system performance is affected by low quality iris images. With the development of electronic color imaging, there are more and more researches about visible wavelength (VW) iris recognition. Compared to the near infrared iris images, using VW iris images acquired under unconstrained imaging conditions is a more challenging task for the iris recognition system. However, the number of quality assessment methods for VW iris images is limited. Therefore, it is interested to investigate whether existing no-reference image quality metrics (IQMs) which are designed for natural images can assess the quality of VW iris images. In this paper, we evaluate the performance of 15 selected no-reference IQMs on VW iris biometrics. The experimental results show that several IQMs can assess iris sample quality according to the system performance.
