Pupil reactions in cataract surgery videos
Our paper has been accepted for publication in the PLOS ONE journal.
Title: Automatic detection of pupil reactions in cataract surgery videos
Authors: Natalia Sokolova, Klaus Schoeffmann, Mario Taschwer, Stephanie Sarny, Doris Putzgruber-Adamitsch, Yosuf El-Shabrawi
Abstract: Nowadays, post-operative analysis of cataract surgeries becomes more and more
important especially to detect intraoperative complications. Some severe complications
may arise from sudden pupil reactions. These may lead to a significant damage of ocular
structure, especially in inexperienced surgeons. Therefore, the automatic retrieval of
such events may be a great support for the post-operative analysis. This helps to train
young surgeons to deal with such situations. In this work, we automatically detect pupil
reactions in cataract surgery videos. We employ the Mask R-CNN architecture as a
segmentation algorithm, which allows us to segment the pupil and iris with pixel-based
accuracy and then track their size changes across the entire video. We can detect pupil
reactions with a harmonic mean (H) of Recall, Precision, and Ground Truth Coverage
Rate (GTCR) of 60, 9% and average prediction length (PL) of 18.93 seconds. However,
we consider the best configuration for practical use the one with an H value of 59.4%
and PL of 10.2 seconds, which is much shorter. We further investigate the
generalization of this method on a slightly different dataset without retraining the
model. In this evaluation, we achieve an H value of 49.3% with the PL of 18.15 seconds.