English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

SlotVTG: Object-Centric Adapter for Generalizable Video Temporal Grounding

Forum topic · 小凯 · 2026-03-28

Summary

SlotVTG (arXiv:2603.25733) is a framework from researchers including Jiwook Han and Jinwoo Choi that improves the out-of-domain generalization of multimodal large language models on Video Temporal Grounding (VTG). While MLLMs perform strongly on VTG, their coarse recognition limits fine-grained temporal understanding, and task-specific fine-tuning causes models to memorize dataset-specific shortcuts instead of grounding in actual visual content, hurting OOD robustness. Existing object-centric learning remedies require re-running the entire multi-stage training pipeline from scratch. SlotVTG instead adds a lightweight slot adapter that uses slot attention to decompose visual tokens into abstract slots and reconstruct the original sequence, with object priors from self-supervised vision models encouraging semantically coherent slot formation. Cross-domain evaluations on standard VTG benchmarks show significantly improved OOD robustness while maintaining competitive in-domain performance, at minimal overhead.

Paper Overview

  • Research field: Computer Vision
  • Authors: Jiwook Han, Geo Ahn, Youngrae Kim, Jinwoo Choi
  • Published: 2026-03-26
  • arXiv: 2603.25733v1
  • Summary

    Multimodal Large Language Models (MLLMs) have shown strong performance on Video Temporal Grounding (VTG). However, their coarse recognition capabilities are insufficient for fine-grained temporal understanding, making task-specific fine-tuning indispensable. This fine-tuning causes models to memorize dataset-specific shortcuts rather than faithfully grounding in the actual visual content, leading to poor Out-of-Domain (OOD) generalization.

    Object-centric learning offers a promising remedy by decomposing scenes into entity-level representations, but existing approaches require re-running the entire multi-stage training pipeline from scratch. The authors propose SlotVTG, a framework that steers MLLMs toward object-centric, input-grounded visual reasoning at minimal cost.

    Method

  • SlotVTG introduces a lightweight slot adapter that applies slot attention to decompose visual tokens into abstract slots and then reconstructs the original sequence.
  • Object priors from self-supervised visual models encourage the formation of semantically coherent slots.
  • The adapter approach avoids costly full multi-stage retraining of the MLLM.
  • Results

    Cross-domain evaluations on standard VTG benchmarks show that SlotVTG:

  • Significantly improves OOD robustness compared to standard fine-tuned MLLMs.
  • Maintains competitive in-domain (ID) performance.
  • Adds only minimal computational overhead.
---

*Auto-collected on 2026-03-28*

Tags

#computer-vision#video-temporal-grounding#mllm#object-centric-learning#slot-attention#out-of-distribution#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/177169367