Unsupervised Instance Segmentation with Superpixels

2025-09-02

AI Summary

To address the heavy reliance of instance segmentation on costly, labor-intensive manual annotations, this paper proposes a fully unsupervised instance segmentation framework. Methodologically, it introduces the first integration of superpixels—generated via MultiCut and low-level features—with self-supervised visual representations, and designs a superpixel-guided mask loss with dual hard and soft branches. Furthermore, an adaptive-weighted self-training mechanism is incorporated to enable pseudo-label quality-driven iterative optimization. The core contributions are: (1) joint modeling of superpixels and self-supervised features; (2) a two-stage learnable mask loss function; and (3) an adaptive self-training strategy. Evaluated on standard benchmarks, the proposed method achieves state-of-the-art performance in both unsupervised instance segmentation and unsupervised object detection, outperforming all existing approaches.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Segmentation
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Instance segmentation is essential for numerous computer vision applications, including robotics, human-computer interaction, and autonomous driving. Currently, popular models bring impressive performance in instance segmentation by training with a large number of human annotations, which are costly to collect. For this reason, we present a new framework that efficiently and effectively segments objects without the need for human annotations. Firstly, a MultiCut algorithm is applied to self-supervised features for coarse mask segmentation. Then, a mask filter is employed to obtain high-quality coarse masks. To train the segmentation network, we compute a novel superpixel-guided mask loss, comprising hard loss and soft loss, with high-quality coarse masks and superpixels segmented from low-level image features. Lastly, a self-training process with a new adaptive loss is proposed to improve the quality of predicted masks. We conduct experiments on public datasets in instance segmentation and object detection to demonstrate the effectiveness of the proposed framework. The results show that the proposed framework outperforms previous state-of-the-art methods.

Problem

• Segments objects without human annotations
• Uses superpixels and self-supervised features
• Improves mask quality through adaptive loss training

Innovation

• Unsupervised segmentation using MultiCut algorithm
• Superpixel-guided mask loss with dual components
• Self-training process with adaptive loss enhancement