Files and links (1)
SubmittedarXiv.org - Non-exclusive license to distribute, Open
Conference proceeding
KeyPoint Relative Position Encoding for Face Recognition
2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 244-255
16 Jun 2024
Abstract
In this paper, we address the challenge of making ViT models more robust to unseen affine transformations. Such robustness becomes useful in various recognition tasks such as face recognition when image alignment failures occur. We propose a novel method called KP-RPE, which leverages key points (e.g. facial landmarks) to make ViT more resilient to scale, translation, and pose variations. We begin with the observation that Relative Position Encoding (RPE) is a good way to bring affine transform generalization to ViTs. RPE, however, can only inject the model with prior knowledge that nearby pixels are more important than far pixels. Keypoint RPE (KP-RPE) is an extension of this principle, where the significance of pixels is not solely dictated by their proximity but also by their relative positions to specific keypoints within the image. By anchoring the significance of pixels around keypoints, the model can more effectively retain spatial relationships, even when those relationships are disrupted by affine transformations. We show the merit of KP-RPE inface and gait recognition. The experimental results demonstrate the effectiveness in improving face recognition performance from low-quality images, particularly where alignment is prone to failure. Code and pre-trained models are available.
Metrics
Details
- Title
- KeyPoint Relative Position Encoding for Face Recognition
- Creators
- Minchul Kim - Michigan State UniversityYiyang Su - Michigan State UniversityFeng Liu - Michigan State UniversityAnil Jain - Michigan State UniversityXiaoming Liu - Michigan State UniversityFeng Liu - Drexel University, Computer Science
- Publication Details
- 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 244-255
- Publisher
- IEEE
- Number of pages
- 12
- Grant note
- Office of the Director of National Intelligence (ODNI) (10.13039/100011038) 2022-21102100004 / Intelligence Advanced Research Projects Activity (IARPA) (10.13039/100011039)
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Computer Science
- Web of Science ID
- WOS:001322555900022
- Scopus ID
- 2-s2.0-85208588876
- Other Identifier
- 991022008296204721
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Web of Science research areas
- Computer Science, Artificial Intelligence
- Computer Science, Interdisciplinary Applications
- Computer Science, Theory & Methods