AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Diff-MM: Exploring Pre-Trained Text-to-Image Generation Models for Unified Multi-Modal Object Tracking

Shiyu Xuan, Zechao Li, Jinhui Tang, Maocheng Zhao

TPAMI 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multi-modality image fusion · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use parallel feature extraction (PFE) instead of Diff-Tracker prompt tokens (ablation)Diff-MM · LaSOTAUC: 58.7 -> 71.5Improved
use PFE instead of cross-attention with MAE template encoder (ablation)Diff-MM · LaSOTAUC: 65.7 -> 71.5Improved
use PFE instead of cross-attention with CLIP ViT template encoder (ablation)Diff-MM · LaSOTAUC: 64.5 -> 71.5Improved
use PFE instead of Diff-Tracker (ablation)Diff-MM · TrackingNetAUC: 75.6 -> 84.7Improved
add auxiliary depth modality via multi-modal sub-module tuning (ablation)Diff-MM · DepthTrackF-score: 58.5 -> 65.3Improved
use sub-module instead of early fusion (ablation)Diff-MM · DepthTrackF-score: 60.9 -> 65.3Improved
reuse VAE encoder instead of T2I-Adapter extra encoder (ablation)Diff-MM · LasHeRPR: 63.7 -> 71.3Improved
freeze UNet in second stage instead of tuning it (ablation)Diff-MM · DepthTrackF-score: 62.3 -> 65.3Improved
add zero-init convolution to lateral connections (ablation)Diff-MM · DepthTrackF-score: 62.6 -> 65.3Improved
add TNL2K to first-stage training data (ablation)Diff-MM · DepthTrackF-score: 65.5 -> 65.3No clear effect
use SD v3-5 (MM-DiT) as base model instead of SD v1-5Diff-MM 256 · TNL2KAUC: 65.2 -> 67.5Improved
use single parameter set instead of modality-specific parametersDiff-MM 256 · DepthTrackF-score: 65.3 -> 65.9No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy