IMAGE DIFFERENCES ANALYSIS USING FINDCONTOURS

One of the common problems in find the difference games is trying to quickly and accurately detect the difference between two pictures. Especially if the image has a similar appearance that is difficult to detect with the naked eye. For example, in a game, many of them have difficulty distinguishing based on color, number of objects, pattern complexity, or the like. This project concentrates on how the system can look for differences in the two images by the color of each pixel. Images taken via the neok12 and hellokids websites will then be checked for differences in each pixel and use the findContours algorithm to extract the contours. The results obtained from this study are 72% success in detecting differences and having an error rate of 28% in 50 trials using 100 images.