Resolution as a Direction: Vector-Panning Feature Alignment for Cross-Resolution Re-Identification

Zanwu Liu, Chao Yuan, Bo Li, Xiaowei Zhang, Guanglin Niu
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:79598-79614, 2026.

Abstract

Cross-resolution person re-identification (CR-ReID) remains challenging in practical surveillance, where camera quality and capture distance lead to substantial resolution gaps between low-resolution (LR) queries and high-resolution (HR) gallery images. Prior approaches commonly rely on super-resolution (SR) or resolution-invariant representation learning, which often increases system complexity and may not directly address the feature mismatch induced by resolution degradation. In this work, we report a new empirical finding from a dedicated analysis in which identity-specific variation is averaged out: the HR–LR feature discrepancy produced by standard ReID backbones exhibits a consistent, resolution-related semantic direction in the embedding space. We further support this observation with statistical analyses based on Canonical Correlation Analysis (CCA) and Pearson correlation analysis. Motivated by this finding, we propose Vector Panning Feature Alignment (VPFA), a lightweight post-hoc module that learns to pan LR features along the learned resolution direction to obtain pseudo-HR representations. VPFA operates after feature extraction and can be integrated into existing ReID systems with negligible overhead. Extensive experiments on multiple CR-ReID benchmarks show that VPFA achieves state-of-the-art performance while improving efficiency compared to SR-based or jointly trained alternatives. Code is available at https://github.com/ashmentlzw/VPFA

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-liu26fr, title = {Resolution as a Direction: Vector-Panning Feature Alignment for Cross-Resolution Re-Identification}, author = {Liu, Zanwu and Yuan, Chao and Li, Bo and Zhang, Xiaowei and Niu, Guanglin}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {79598--79614}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/liu26fr/liu26fr.pdf}, url = {https://proceedings.mlr.press/v306/liu26fr.html}, abstract = {Cross-resolution person re-identification (CR-ReID) remains challenging in practical surveillance, where camera quality and capture distance lead to substantial resolution gaps between low-resolution (LR) queries and high-resolution (HR) gallery images. Prior approaches commonly rely on super-resolution (SR) or resolution-invariant representation learning, which often increases system complexity and may not directly address the feature mismatch induced by resolution degradation. In this work, we report a new empirical finding from a dedicated analysis in which identity-specific variation is averaged out: the HR–LR feature discrepancy produced by standard ReID backbones exhibits a consistent, resolution-related semantic direction in the embedding space. We further support this observation with statistical analyses based on Canonical Correlation Analysis (CCA) and Pearson correlation analysis. Motivated by this finding, we propose Vector Panning Feature Alignment (VPFA), a lightweight post-hoc module that learns to pan LR features along the learned resolution direction to obtain pseudo-HR representations. VPFA operates after feature extraction and can be integrated into existing ReID systems with negligible overhead. Extensive experiments on multiple CR-ReID benchmarks show that VPFA achieves state-of-the-art performance while improving efficiency compared to SR-based or jointly trained alternatives. Code is available at https://github.com/ashmentlzw/VPFA} }
Endnote
%0 Conference Paper %T Resolution as a Direction: Vector-Panning Feature Alignment for Cross-Resolution Re-Identification %A Zanwu Liu %A Chao Yuan %A Bo Li %A Xiaowei Zhang %A Guanglin Niu %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-liu26fr %I PMLR %P 79598--79614 %U https://proceedings.mlr.press/v306/liu26fr.html %V 306 %X Cross-resolution person re-identification (CR-ReID) remains challenging in practical surveillance, where camera quality and capture distance lead to substantial resolution gaps between low-resolution (LR) queries and high-resolution (HR) gallery images. Prior approaches commonly rely on super-resolution (SR) or resolution-invariant representation learning, which often increases system complexity and may not directly address the feature mismatch induced by resolution degradation. In this work, we report a new empirical finding from a dedicated analysis in which identity-specific variation is averaged out: the HR–LR feature discrepancy produced by standard ReID backbones exhibits a consistent, resolution-related semantic direction in the embedding space. We further support this observation with statistical analyses based on Canonical Correlation Analysis (CCA) and Pearson correlation analysis. Motivated by this finding, we propose Vector Panning Feature Alignment (VPFA), a lightweight post-hoc module that learns to pan LR features along the learned resolution direction to obtain pseudo-HR representations. VPFA operates after feature extraction and can be integrated into existing ReID systems with negligible overhead. Extensive experiments on multiple CR-ReID benchmarks show that VPFA achieves state-of-the-art performance while improving efficiency compared to SR-based or jointly trained alternatives. Code is available at https://github.com/ashmentlzw/VPFA
APA
Liu, Z., Yuan, C., Li, B., Zhang, X. & Niu, G.. (2026). Resolution as a Direction: Vector-Panning Feature Alignment for Cross-Resolution Re-Identification. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:79598-79614 Available from https://proceedings.mlr.press/v306/liu26fr.html.

Related Material