Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

HyperVLP: Enhancing Hierarchical Surgical Video-Language Pre-training in Hyperbolic Space

About

Surgical vision-language foundation models typically adopt educational materials, such as surgical lecture videos, to transfer surgical knowledge encoded in language into visual representations. These knowledge are multi-dimensional and hierarchical: fine-grained action cues appear in narration, mid-level key steps are summarized in subsection headings, and global procedural context, such as patient history and surgical strategy, is described in abstract texts. Prior work largely collapses these heterogeneous signals into a single flat embedding space, implicitly assuming independence across hierarchy levels. However, this is suboptimal because it ignores cross-level semantic containment, e.g., actions belong to steps, steps compose phases, weakens long-range dependency modeling. To this end, we propose a hyperbolic surgical video-language pre-training framework that explicitly preserves the hierarchical structure by mitigating structural false negatives induced by procedural context and enforcing semantic consistency between parent phases and their constituent child steps. Extensive experiments on multiple surgical benchmarks show consistent gains in zero- and few-shot phase recognition across procedures and institutions.

Yaojun Hu, Kun Yuan, Nassir Navab, Haochao Ying, Jian Wu, Nicolas Padoy• 2026

Related benchmarks

TaskDatasetResultRank
Surgical Phase RecognitionCholec80
Accuracy71.3
70
Phase RecognitionCholec80 (test)
F1 Score0.372
37
Phase RecognitionAutoLaparo (test)
F1 Score32.9
30
Phase RecognitionStrasBypass70 (test)
Top-1 Accuracy52.7
21
Phase RecognitionBernBypass70 (test)
Top-1 Accuracy50
21
Surgical Phase RecognitionAutolaparo 100%
Accuracy65.3
9
Surgical Phase RecognitionAutolaparo (10%)
Accuracy60.1
9
Surgical Phase RecognitionStras70 (100%)
Accuracy75.6
9
Surgical Phase RecognitionStras70 (10%)
Accuracy66.9
9
Surgical Phase RecognitionBern70 (100%)
Accuracy76.9
9
Showing 10 of 12 rows

Other info

Follow for update