The project is aimed at improving the accuracy of the detection of various phases involved in the cholecystectomy surgery (surgical removal of the gallbladder). For this purpose, we evaluated various Neural network architectures. The automated surgical workflow analysis can have multiple applications in training surgeons and standardization of medical procedures.
Automated Surgical workflow analysis has been given a lot of importance lately to keep up with the demands of the modern healthcare industry and the significant benefits associated with the same. The automated surgical workflow phase detection forms an important subpart of the analysis, and has several medical applications, such as automated indexing of surgical video databases, grading a new surgeon based on the data from the surgery performed by the experts, etc. A large portion of a surgeon’s time is spent in learning and subsequently training further surgeons in such medical procedures, which can be used upon. It can also help surgeons to standardize procedures and enhance post - surgical assessment. The accuracy improvement of the phase-detection can ultimately lead to higher dependence on such an automated procedure. This objective evaluation can be compared to an expert surgeons' assessment on a video-enabled surgical procedure. Developing countries with limited trained manpower and resources can heavily rely on this automated procedure to cater to their medical needs.
Computer Vision has found its use in every domain and healthcare is not left behind. Not only healthcare has been made smarter by Computer Vision and Machine Learning techniques, but they have proved to have life-impacting potential. This project is aimed at improving the accuracy of the detection of various phases involved in the cholecystectomy surgery (surgical removal of the gallbladder) using Computer Vision techniques. Every year, more than 3 lac such surgeries are performed in the United States alone, which further strengthens the motivation to work on such a project.
For such an automated analysis, accuracy improvement has been given the foremost priority, as it can have impacts on multiple lives. The training time, though an important factor, has not been given an equal priority considering the far -reaching consequences, a wide number of applications and a general impact on the further training of inexperienced doctors.