September 1, 2024Revista Mexicana de Ingeniería Biomédica

UMInSe: An Unsupervised Method for Segmentation and Detection of Surgical Instruments based on K-means

View Full Paper

Ask AI

Bookmark

Accurate surgical instrument segmentation is achieved without manual annotations, enabling automated neural network training across varying illumination and occlusions.
High accuracy of 0.9602 and a minimal error of 0.0297 confirm effective instrument segmentation across surgical benchmark datasets, demonstrating unsupervised machine learning robustness.
Assessment using the JIGSAWS and EndoVis databases applies Gamma correction to identify regions of interest, supporting automated robot motion planning during complex surgeries.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Benchmark evaluation demonstrates accurate surgical instrument segmentation in endoscopic surgical datasets, highlighting potential for automated tool tracking without manual labels.

Cite This Study

Arevalo-Ancona et al. (2024) studied this question.

synapsesocial.com/papers/68e59fa4b6db643587539e66 https://doi.org/10.17488/rmib.45.3.2

View Full Paper

Ask AI

Bookmark

Implication

Benchmark evaluation demonstrates accurate surgical instrument segmentation in endoscopic surgical datasets, highlighting potential for automated tool tracking without manual labels.

UMInSe: An Unsupervised Method for Segmentation and Detection of Surgical Instruments based on K-means

Key Points

Discussion

Implication

Cite This Study

UMInSe: An Unsupervised Method for Segmentation and Detection of Surgical Instruments based on K-means

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider