Home >

news Help

Publication Information


Title
Japanese: 
English:Diffusion Pretraining for Gait Recognition in the Wild 
Author
Japanese: Wei Ming Neo, 篠田浩一, Tat-Jen Cham.  
English: Wei Ming Neo, Koichi Shinoda, Tat-Jen Cham.  
Language English 
Journal/Book name
Japanese: 
English:2025 IEEE International Conference on Image Processing (ICIP) 
Volume, Number, Page         pp. 1295 - 1300
Published date Sept. 14, 2025 
Publisher
Japanese: 
English:IEEE 
Conference name
Japanese: 
English:The IEEE International Conference on Image Processing (ICIP) 2025 
Conference site
Japanese: 
English:Alaska, Anchorage 
File
DOI https://doi.org/10.1109/ICIP55913.2025.11084665
Abstract Recently, diffusion models have garnered much attention for their remarkable generative capabilities. Yet, their application for representation learning remains largely unexplored. In this paper, we explore the potential of diffusion models to pretrain the backbone of a deep learning model for a specific application—gait recognition in the wild. To do so, we condition a latent diffusion model on the output of a gait recognition model backbone. Our pretraining experiments on the Gait3D and GREW datasets reveal an interesting phenomenon: diffusion pretraining causes the gait recognition backbone to separate gait sequences belonging to different subjects further apart than those belonging to the same subjects. Subsequently, our transfer learning experiments on Gait3D and GREW show that the pretrained backbone can serve as an effective initialization for the downstream gait recognition task, improving gait recognition accuracies by as much as 7.9% on Gait3D and 4.2% on GREW.

©2007 Institute of Science Tokyo All rights reserved.