EVENT DOSSIER
The research team proposed the WeakMCN multi-task collaboration network, aimed at improving both the performance in weak supervision representation understanding (WREC) and segmentation (WRES). This network adopts a two-branch architecture, where the WREC branch is based on contrastive learning with anchors and supervises the WRES branch. Innovatively, dynamic visual feature enhancement (DVFE) and collaborative consistency module (CCM) were introduced. Experiments showed that on the RefCOCO, RefCOCO+, and RefCOCOg benchmarks, the performance in WREC and WRES increased by 3.91% and 13.11%, respectively. Additionally, under semi-supervised settings, on RefCOCO with only 1% labeled data, the performance in semi-symbolic understanding and segmentation increased by 8.94% and 7.71%, respectively, verifying its strong generalization ability in weak supervision scenarios and its superior performance compared to existing single-task methods.
arXiv:2505.18686v3 WeakMCN 提出多任务协作网络，提升弱监督指代理解与分割性能
提出 WeakMCN 多任务协作网络，通过双分支架构结合弱监督指称表达理解（WREC）与分割（WRES），在 RefCOCO、RefCOCO+ 及 RefCOCOg 三个基准上分别提升 WREC 和 WRES 性能达 3.91% 和 13.11%，并在半监督设置下使 1% RefCOCO 上的半指称理解与分割性能分别提升 8.94% 和 7.71%。该网络采用基于锚点的对比学习作为 WREC 分支并监督 WRES 分支，引入动态视觉特征增强（DVFE）与协作一致性模块（CCM）两项创新设计以优化多任务协作，实验验证了其在现有最先进单任务方法上的显著优势及在弱监督场景下的强泛化能力。