.png){kind=icon}
.png){kind=icon}
Return
Weak-edge sample extension for enhancing unsupervised feature learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.131171.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• An unsupervised Re-ID framework based on sample extension is proposed to learn discriminative features from weak-edge samples. • An edge strength scoring mechanism is proposed to capture neighborhood structure information within the cluster to estimate sample positions. • An edge-aware sample extension module based on the edge strength score is proposed to provide tighter structural support for weak-edge samples, enabling the model to learn discriminative features better. • Extensive experimental results demonstrate that the proposed method achieves competitive performance compared to the best existing approaches.
Keywords:
unsupervised Re-ID
sample extension
edge strength scoring
discriminative features
weak-edge samples