CatStep: Automated Cataract Surgical Phase Classification and Boundary Segmentation Leveraging Inflated 3D-Convolutional Neural Network Architectures and BigCat

Review Engineering, Biomedical

Large-scale surgical workflow segmentation for laparoscopic sacrocolpopexy

Yitong Zhang et al.

Summary: This study investigates the workflow segmentation of sacrocolpopexy videos and finds that it differs from the commonly used benchmark Cholec80, requiring dedicated approaches due to longer phase durations. Sequence-to-sequence models are used for coarse-level phase segmentation, and different architectures and configurations are tested and evaluated. The results show a trade-off between accurate frames and consistent phase transitions for different architectures. Additionally, an event-based evaluation metric proves useful for evaluating workflow segmentation algorithms.

INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY (2022)

Add to Collection

Article Ophthalmology

Evaluation of Artificial Intelligence-Based Intraoperative Guidance Tools for Phacoemulsification Cataract Surgery

Rogerio Garcia Nespolo et al.

Summary: This study evaluated the use of a deep neural network to track the pupil, identify the surgical phase, and provide visual feedback in real time during phacoemulsification cataract surgery. The results showed that the neural network model achieved high accuracy in identifying the surgical phase and pupil segmentation, and the surgeons surveyed expressed a high likelihood of using the platform. These findings suggest that an artificial intelligence-based surgical guidance platform can enhance the surgical experience in phacoemulsification cataract surgery.

JAMA OPHTHALMOLOGY (2022)

Add to Collection

Article Ophthalmology

Analysis of Cataract Surgery Instrument Identification Performance of Convolutional and Recurrent Neural Network Ensembles Leveraging BigCat

Nicholas Matton et al.

Summary: This study developed a method for accurate automated real-time identification of instruments in cataract surgery videos. By utilizing a dense convolutional neural network and a large dataset, the authors achieved state-of-the-art results in surgical instrument identification.

TRANSLATIONAL VISION SCIENCE & TECHNOLOGY (2022)

Add to Collection

Article Computer Science, Artificial Intelligence