English

Multi-Modal Representation Learning with Text-Driven Soft Masks

Computer Vision and Pattern Recognition 2023-04-04 v1

Abstract

We propose a visual-linguistic representation learning approach within a self-supervised learning framework by introducing a new operation, loss, and data augmentation strategy. First, we generate diverse features for the image-text matching (ITM) task via soft-masking the regions in an image, which are most relevant to a certain word in the corresponding caption, instead of completely removing them. Since our framework relies only on image-caption pairs with no fine-grained annotations, we identify the relevant regions to each word by computing the word-conditional visual attention using multi-modal encoder. Second, we encourage the model to focus more on hard but diverse examples by proposing a focal loss for the image-text contrastive learning (ITC) objective, which alleviates the inherent limitations of overfitting and bias issues. Last, we perform multi-modal data augmentations for self-supervised learning via mining various examples by masking texts and rendering distortions on images. We show that the combination of these three innovations is effective for learning a pretrained model, leading to outstanding performance on multiple vision-language downstream tasks.

Keywords

Cite

@article{arxiv.2304.00719,
  title  = {Multi-Modal Representation Learning with Text-Driven Soft Masks},
  author = {Jaeyoo Park and Bohyung Han},
  journal= {arXiv preprint arXiv:2304.00719},
  year   = {2023}
}

Comments

CVPR 2023

R2 v1 2026-06-28T09:45:48.275Z