[AAAI'26] ReTrack: Evidence-Driven Dual-Stream Directional Anchor Calibration Network for Composed Video Retrieval

Zixu Li1, Yupeng Hu1*, Zhiwei Chen1, Qinlei Huang1, Guozhi Qiu1, Zhiheng Fu1, Meng Liu2,
1School of Software, Shandong University,
2School of Computer Science and Technology, Shandong Jianzhu University
*Corresponding author.
Paper Complete source code

Abstract

MY ALT TEXT

Illustration of Directional Bias

MY ALT TEXT

(a) illustrates a typical CVR example. (b) highlights the directional bias issue in existing methods, where the similarity between the composed feature and the target video becomes indistinguishable from that of certain negative candidates, degrading retrieval performance. (c) demonstrates that our method effectively mitigates directional bias, producing a clear separation between the composed feature's similarity to the target and all negative samples.


Framework: evidence- dRivEn dual-sTream diRectionAl anChor calibration networK (ReTrack)

MY ALT TEXT

The proposed ReTrack consists of three key modules: (a) Semantic Contribution Disentanglement, (b) Composition Geometry Calibration, and (c) Reliable Evidence-driven Alignment.

Experiment

MY ALT TEXT

Performance comparison on the test set of the CVR dataset, WebVid-CoVR, relative to R@k(%). The overall best results are in bold, while the best results over baselines are underlined.


MY ALT TEXT

Performance comparison on the CIR dataset, FashionIQ and CIRR, relative to R@k(%). The overall best results are in bold, while the best results over baselines are underlined.


MY ALT TEXT

Ablation study on three CVR and CIR datasets.


MY ALT TEXT

Sensitivity to the hyper-parameters (a) κ, and (b) λ on WebVid-CoVR and CIRR datasets.


MY ALT TEXT

Case study on (a) WebVid-CoVR and (b) CIRR.

BibTeX


        @inproceedings{ReTrack,
          title={ReTrack: Evidence Driven Dual Stream Directional Anchor Calibration Network for Composed Video Retrieval},
          author={Li, Zixu and Hu, Yupeng and Chen, Zhiwei and Huang, Qinlei and Qiu, Guozhi and Fu, Zhiheng and Liu, Meng},
          booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
          year={2026}
        }