Betrayed by Captions: Joint Caption Grounding and Generation for Open Vocabulary Instance Segmentation

TL;DR

This paper introduces a joint caption grounding and generation framework that improves open vocabulary instance segmentation by effectively associating image regions with novel object categories, outperforming previous methods.

Contribution

The proposed CGG framework combines a novel grounding loss with caption generation to enhance learning and contextual understanding for segmenting unseen object categories.

Findings

Achieves 6.8% mAP improvement for novel classes on OVIS.

Attains 15% PQ improvement for novel classes on OSPS.

Demonstrates the effectiveness of joint grounding and generation in open vocabulary segmentation.

Abstract

In this work, we focus on open vocabulary instance segmentation to expand a segmentation model to classify and segment instance-level novel categories. Previous approaches have relied on massive caption datasets and complex pipelines to establish one-to-one mappings between image regions and words in captions. However, such methods build noisy supervision by matching non-visible words to image regions, such as adjectives and verbs. Meanwhile, context words are also important for inferring the existence of novel objects as they show high inter-correlations with novel categories. To overcome these limitations, we devise a joint \textbf{Caption Grounding and Generation (CGG)} framework, which incorporates a novel grounding loss that only focuses on matching object nouns to improve learning efficiency. We also introduce a caption generation head that enables additional supervision and contextual modeling as a complementation to the grounding loss. Our analysis and results demonstrate that grounding and generation components complement each other, significantly enhancing the segmentation performance for novel classes. Experiments on the COCO dataset with two settings: Open Vocabulary Instance Segmentation (OVIS) and Open Set Panoptic Segmentation (OSPS) demonstrate the superiority of the CGG. Specifically, CGG achieves a substantial improvement of 6.8% mAP for novel classes without extra data on the OVIS task and 15% PQ improvements for novel classes on the OSPS benchmark.