University of Twente Student Theses
Temporal Action Segmentation in Laparoscopic Surgery Videos : An Evaluation Study
Gurev, Crina (2025) Temporal Action Segmentation in Laparoscopic Surgery Videos : An Evaluation Study.
This is the latest version of this item.
|
PDF
3MB |
| Abstract: | Temporal Action Segmentation (TAS) is a crucial task in video understanding, aimed at segmenting videos into distinct temporal actions and assigning pre-defined labels to each segment. This study addresses the challenges of TAS in laparoscopic surgery videos, characterized by lack of data, dynamic transitions and non-static backgrounds, by proposing an unsupervised framework. We utilize ResNet-101 for feature extraction, focusing on both low-level features (e.g., edges, textures) and high-level features (e.g., object semantics, spatial relationships). These features are used to evaluate and compare the clustering performance of two widely used algorithms: Normalized Spectral Clustering (NSC) and Agglomerative Hierarchical Clustering (AHC). The framework is validated on a custom-annotated dataset of laparoscopic surgery videos, using both frame-level and boundary-detection evaluation metrics such as Precision, Recall and F1-Score. This research aims to provide insights into the effectiveness of NSC versus AHC and the impact of low-level versus high-level features in accurately segmenting complex surgical videos, offering a valuable contribution to medical video analysis and training. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/105133 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page