PulseAugur
EN
LIVE 11:46:50

New CRST framework improves low-resolution text-to-image person retrieval

Researchers have developed a new framework called Cross-Resolution Semantic Transfer (CRST) to improve text-to-image person re-identification, particularly in low-resolution surveillance scenarios. CRST addresses issues like unreliable evidence from degraded images and distorted retrieval rankings caused by mixed resolutions. The framework utilizes resolution-conditioned reasoning, text-guided refinement, and a novel CR-RDA module to enhance retrieval accuracy and stability across various resolutions. AI

IMPACT This research could lead to more robust person identification systems in surveillance, even with low-quality imagery.

RANK_REASON The cluster contains a research paper detailing a new framework for text-to-image retrieval.

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New CRST framework improves low-resolution text-to-image person retrieval

COVERAGE [2]

  1. arXiv cs.CV TIER_1 English(EN) · Wenjie Qian, Bin Yang, Xiao Wang, Wenke Huang, Ling Mei, Xin Xu, Mang Ye ·

    Cross-Resolution Semantic Transfer for Robust Text-to-Image Retrieval in Low-Resolution Surveillance

    arXiv:2606.30458v1 Announce Type: new Abstract: Text-to-image person re-identification (TIPR) retrieves target persons using natural language descriptions. However, existing methods largely overlook resolution variance in real-world surveillance. They characterize cross-resolutio…

  2. arXiv cs.CV TIER_1 English(EN) · Mang Ye ·

    Cross-Resolution Semantic Transfer for Robust Text-to-Image Retrieval in Low-Resolution Surveillance

    Text-to-image person re-identification (TIPR) retrieves target persons using natural language descriptions. However, existing methods largely overlook resolution variance in real-world surveillance. They characterize cross-resolution TIPR through two coupled failure modes: Eviden…