IntRec: Intent-based Retrieval with Contrastive Refinement

TL;DR

IntRec is an interactive object retrieval framework that uses user feedback and contrastive refinement to improve accuracy in complex scenes, outperforming existing methods with minimal latency.

Contribution

We introduce IntRec, a novel interactive retrieval method that maintains dual memory sets and employs contrastive alignment for fine-grained disambiguation based on user feedback.

Findings

Achieves 35.4 AP on LVIS, surpassing competitors.

Improves by +7.9 AP on LVIS-Ambiguous after one feedback.

Operates with less than 30 ms latency per interaction.

Abstract

Retrieving user-specified objects from complex scenes remains a challenging task, especially when queries are ambiguous or involve multiple similar objects. Existing open-vocabulary detectors operate in a one-shot manner, lacking the ability to refine predictions based on user feedback. To address this, we propose IntRec, an interactive object retrieval framework that refines predictions based on user feedback. At its core is an Intent State (IS) that maintains dual memory sets for positive anchors (confirmed cues) and negative constraints (rejected hypotheses). A contrastive alignment function ranks candidate objects by maximizing similarity to positive cues while penalizing rejected ones, enabling fine-grained disambiguation in cluttered scenes. Our interactive framework provides substantial improvements in retrieval accuracy without additional supervision. On LVIS, IntRec achieves 35.4 AP, outperforming OVMR, CoDet, and CAKE by +2.3, +3.7, and +0.5, respectively. On the challenging LVIS-Ambiguous benchmark, it improves performance by +7.9 AP over its one-shot baseline after a single corrective feedback, with less than 30 ms of added latency per interaction.