Slot-BERT: Self-supervised Object Discovery in Surgical Video

TL;DR

Slot-BERT is a novel bidirectional model that learns object-centric representations in surgical videos, maintaining long-range temporal coherence efficiently and enabling unsupervised discovery and domain adaptation across diverse surgical procedures.

Contribution

It introduces Slot-BERT, a scalable, bidirectional, self-supervised model with a novel contrastive loss for object discovery in long surgical videos, outperforming existing methods.

Findings

Outperforms state-of-the-art object-centric methods in surgical video analysis.

Achieves robust zero-shot domain adaptation across different surgical datasets.

Scales effectively to long, unconstrained surgical videos.

Abstract

Object-centric slot attention is a powerful framework for unsupervised learning of structured and explainable representations that can support reasoning about objects and actions, including in surgical videos. While conventional object-centric methods for videos leverage recurrent processing to achieve efficiency, they often struggle with maintaining long-range temporal coherence required for long videos in surgical applications. On the other hand, fully parallel processing of entire videos enhances temporal consistency but introduces significant computational overhead, making it impractical for implementation on hardware in medical facilities. We present Slot-BERT, a bidirectional long-range model that learns object-centric representations in a latent space while ensuring robust temporal coherence. Slot-BERT scales object discovery seamlessly to long videos of unconstrained lengths. A novel slot contrastive loss further reduces redundancy and improves the representation disentanglement by enhancing slot orthogonality. We evaluate Slot-BERT on real-world surgical video datasets from abdominal, cholecystectomy, and thoracic procedures. Our method surpasses state-of-the-art object-centric approaches under unsupervised training achieving superior performance across diverse domains. We also demonstrate efficient zero-shot domain adaptation to data from diverse surgical specialties and databases.