Multiple Object Stitching for Unsupervised Representation Learning
This addresses the challenge of learning detailed representations for multi-object images in unsupervised learning, which is crucial for downstream tasks like object detection and semantic segmentation, though it is incremental as it builds on existing contrastive methods.
The paper tackles the problem of inferior unsupervised representation learning for images with multiple objects by proposing Multiple Object Stitching (MOS), which synthesizes multi-object images from single-object ones to provide additional object correspondences without human annotations, achieving leading performance on datasets like ImageNet, CIFAR, and COCO.
Contrastive learning for single object centric images has achieved remarkable progress on unsupervised representation, but suffering inferior performance on the widespread images with multiple objects. In this paper, we propose a simple but effective method, Multiple Object Stitching (MOS), to refine the unsupervised representation for multi-object images. Specifically, we construct the multi-object images by stitching the single object centric ones, where the objects in the synthesized multi-object images are predetermined. Hence, compared to the existing contrastive methods, our method provides additional object correspondences between multi-object images without human annotations. In this manner, our method pays more attention to the representations of each object in multi-object image, thus providing more detailed representations for complicated downstream tasks, such as object detection and semantic segmentation. Experimental results on ImageNet, CIFAR and COCO datasets demonstrate that our proposed method achieves the leading unsupervised representation performance on both single object centric images and multi-object ones. The source code is available at https://github.com/visresearch/MultipleObjectStitching.