Robotizing flexible endoscopy enables image-based control of endoscopes. Especially during high-throughput procedures, such as a colonoscopy, navigation support algorithms could improve procedure turnaround and ergonomics for the endoscopist.In this study, we have developed and implemented a navigation algorithm that is based on image classification followed by dark region segmentation. Robustness and accuracy were evaluated on real images obtained from human colonoscopy exams. Comparison was done using manual annotation as a reference. Intraclass correlation (ICC) was employed as a measure for similarity between automated and manual results.The discrimination of the developed classifier was 6.8, making it a reliable classifier. In the experiments, the developed algorithm gave an ICC of 93 % (range 84.7-98.8 %) over the test image sequences on average. If images were classified as 'uninformative', which led to re-initialization of the algorithm, this was predictive for the result of dark region segmentation accuracy.In conclusion, the developed target detection algorithm provided accurate results and is thought to provide reliable assistance in the clinic. The clinical relevance of this kind of navigation and control is currently being investigated.
Colon cancer screening remains a time-consuming and expensive clinical process. Automating flexible endoscopy has the potential to increase screening efficiency. In this research the images captured by the camera at the endoscope tip are used to find the heading direction of the endoscope. Comparing the current heading direction to the desired target direction in a computer algorithm is expected to allow automated steering of the endoscope. Heading direction determination is achieved using an estimation of the focus of expansion (FOE) from the optical flow field. The resulting heading direction is compared to results obtained manually by several human observers. From our experiments it becomes clear that the FOE can be used as a reliable estimator for heading direction in human colonoscopy images. Additionally, the automated results have an intraclass correlation of 89% with the manual results, demonstrating that the algorithm works as expected. It is anticipated that the final steering algorithm can be used in a variety of motorized flexible endoscope applications.
I. INTRODUCTIONLEXIBLE endoscopy can be performed to obtain a diagnosis and to perform small interventions in the human body without leaving scars. In screening for colorectal cancer [1], flexible endoscopy of the large bowel (colonoscopy) is performed to diagnose and, if necessary, remove lesions of the bowel wall. It is expected that the number of colonoscopies performed yearly in the Netherlands will increase by 60.000 after the start of the population screening program in 2013 [2]. This rise motivates hospitals to increase their colonoscopy capacity and efficiency drastically.Currently, an endoscope driver (endoscopist) needs to perform a large number of procedures to reach the level of competence. Estimates of this number vary from 100 to 500 colonoscopic procedures over periods of 1 to 3 years Manuscript
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.