CVAILGJun 14, 2022

ReCo: Retrieve and Co-segment for Zero-shot Transfer

CambridgeOxford
arXiv:2206.07045v1147 citationsh-index: 50
Originality Incremental advance
AI Analysis

This work addresses the high annotation costs in semantic segmentation for real-world applications by enabling zero-shot transfer without labeled data, though it builds incrementally on existing retrieval and co-segmentation techniques.

The paper tackled the problem of semantic segmentation without pixel-level supervision by combining retrieval from a language-image pre-trained model (CLIP) with co-segmentation to create a zero-shot transfer method, achieving performance comparable to unsupervised segmentation approaches and enabling segmentation of rare objects.

Semantic segmentation has a broad range of applications, but its real-world impact has been significantly limited by the prohibitive annotation costs necessary to enable deployment. Segmentation methods that forgo supervision can side-step these costs, but exhibit the inconvenient requirement to provide labelled examples from the target distribution to assign concept names to predictions. An alternative line of work in language-image pre-training has recently demonstrated the potential to produce models that can both assign names across large vocabularies of concepts and enable zero-shot transfer for classification, but do not demonstrate commensurate segmentation abilities. In this work, we strive to achieve a synthesis of these two approaches that combines their strengths. We leverage the retrieval abilities of one such language-image pre-trained model, CLIP, to dynamically curate training sets from unlabelled images for arbitrary collections of concept names, and leverage the robust correspondences offered by modern image representations to co-segment entities among the resulting collections. The synthetic segment collections are then employed to construct a segmentation model (without requiring pixel labels) whose knowledge of concepts is inherited from the scalable pre-training process of CLIP. We demonstrate that our approach, termed Retrieve and Co-segment (ReCo) performs favourably to unsupervised segmentation approaches while inheriting the convenience of nameable predictions and zero-shot transfer. We also demonstrate ReCo's ability to generate specialist segmenters for extremely rare objects.

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes