Endoscopic Image Observation Support System for Capsule Endoscopy

The system performs preliminary actions by automatically analyzing images before they reach the interpreter. It calculates moving speeds between sequential images, identifies images where the capsule is stationary or moving slowly, and pre-marks these as potential images of interest. This preliminary processing filters out rapidly moving images that are unlikely to contain diagnostic value, thereby reducing the total number of images requiring manual review while ensuring that all potentially important images are captured for observation.

The system applies local quality by differentiating display time allocation based on image characteristics. Images identified as having the capsule stationary or moving slowly are assigned longer display times to allow thorough interpretation, while images showing rapid movement are assigned shorter display times or excluded from manual review. This selective approach ensures that interpretation accuracy is maintained for images most likely to contain diagnostic information while reducing total observation time by not spending excessive time on images unlikely to be diagnostic.

The system performs preliminary filtering and prioritization before double interpretation begins. It automatically identifies and flags images where the capsule is stationary or moving slowly, creating a prioritized list of images most likely to contain diagnostic information. This preliminary action allows the first interpreter to focus on high-priority images, and enables the second interpreter to efficiently review only necessary cases, thereby maintaining diagnostic accuracy through double interpretation while reducing overall workload and time consumption compared to reviewing all images twice.