Image Segmentation Using Text and Image Prompts

TL;DR

This paper introduces a unified image segmentation system that uses text and image prompts at test time, enabling flexible, multi-task segmentation without retraining for new classes or complex queries.

Contribution

It presents a novel model extending CLIP with a transformer decoder to perform diverse segmentation tasks using arbitrary prompts, trained once on an extended dataset.

Findings

Effective in referring expression, zero-shot, and one-shot segmentation

Handles arbitrary binary segmentation queries with text or image prompts

Shows good adaptation to generalized queries involving properties

Abstract

Image segmentation is usually addressed by training a model for a fixed set of object classes. Incorporating additional classes or more complex queries later is expensive as it requires re-training the model on a dataset that encompasses these expressions. Here we propose a system that can generate image segmentations based on arbitrary prompts at test time. A prompt can be either a text or an image. This approach enables us to create a unified model (trained once) for three common segmentation tasks, which come with distinct challenges: referring expression segmentation, zero-shot segmentation and one-shot segmentation. We build upon the CLIP model as a backbone which we extend with a transformer-based decoder that enables dense prediction. After training on an extended version of the PhraseCut dataset, our system generates a binary segmentation map for an image based on a free-text prompt or on an additional image expressing the query. We analyze different variants of the latter image-based prompts in detail. This novel hybrid input allows for dynamic adaptation not only to the three segmentation tasks mentioned above, but to any binary segmentation task where a text or image query can be formulated. Finally, we find our system to adapt well to generalized queries involving affordances or properties. Code is available at https://eckerlab.org/code/clipseg.