New framework PEC-CIR enhances zero-shot image retrieval with planning and self-criticism

Researchers have developed PEC-CIR, a novel framework designed to improve zero-shot composed image retrieval. This method structures the query construction process into a multi-stage reasoning pipeline, involving a Planner, Executor, and Critic. The Planner identifies explicit constraints, the Executor generates potential target descriptions, and the Critic evaluates these candidates for compliance. By breaking down query generation into these distinct steps and incorporating a self-criticism mechanism, PEC-CIR aims to reduce errors and enhance retrieval stability compared to single-pass generation approaches. AI

IMPACT This research introduces a novel approach to image retrieval that could improve the accuracy and robustness of systems relying on textual descriptions and reference images.

RANK_REASON The item is a research paper detailing a new framework for image retrieval. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• PEC-CIR
• ScienceCast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →