Learning to Prompt Segmentation Foundation Models

Abstract

Segmentation foundation models (SFMs) like SEEM and SAM have demonstrated great potential in learning to segment anything. The core design of SFMs lies with “Promptable Segmentation”, which takes a handcrafted prompt as input and returns the expected segmentation mask. SFMs work with two types of prompts including spatial prompts (e.g., points) and semantic prompts (e.g., texts), which work together to prompt SFMs to segment anything on downstream datasets. Despite the important role of prompts, how to acquire suitable prompts for SFMs is largely under-explored. In this work, we examine the architecture of SFMs and identify two challenges for learning effective prompts for SFMs. To this end, we propose spatial-semantic prompt learning (SSPrompt) that learns effective semantic and spatial prompts for better SFMs. Specifically, SSPrompt introduces spatial prompt learning and semantic prompt learning, which optimize spatial prompts and semantic prompts directly over the embedding space and selectively leverage the knowledge encoded in pre-trained prompt encoders. Extensive experiments show that SSPrompt achieves superior image segmentation performance consistently across multiple widely adopted datasets.