Real-time multilevel sequencing of cataract surgery videos

Data recorded and stored during video-monitored surgeries are a relevant source of information for surgeons, especially during their training period. But today, this data is virtually unexploited. In this paper, we propose to reuse videos recorded during cataract surgeries to automatically analyze the surgical process with the real-time constraint, with the aim to assist the surgeon during the surgery. We propose to automatically recognize, in real-time, what the surgeon is doing: what surgical phase or, more precisely, what surgical step he or she is performing. This recognition relies on the inference of a multilevel statistical model which uses 1) the conditional relations between levels of description (steps and phases) and 2) the temporal relations among steps and among phases. The model accepts two types of inputs: 1) the presence of surgical instruments, manually provided by the surgeons, or 2) motion in videos, automatically analyzed through the CBVR paradigm. A dataset of 30 cataract surgery videos was collected at Brest University hospital. The system was evaluated in terms of mean area under the ROC curve. Promising results were obtained using either motion analysis (Az = 0.759) or the presence of surgical instruments (Az = 0.983).