Comparison of machine and human vision based on brightness and contrast using YOLOV3 with fuzzy logic

This study aims to compare human and machine vision using fuzzy logic with YOLOV3 algorithm based on brightness and contrast values in the images. In order to determine detection thresholds for the contrast and the brightness parameters, we conducted a survey including 50 participants. We also examined how these two parameters affect each other depending on the thresholds. The thresholds are normalized, before being given as input to fuzzy logic. We also analyzed the effect of the weight parameter, that represents the size of the object found in the image. In fact, weight affects the sensitivity of the object in the image to the brightness and contrast parameters. We obtain a fuzzy result by processing the brightness, contrast, and weight parameters with fuzzy logic. This result is compared with the value from YOLOV3. In this way, human and machine vision are compared on the same platform. Depending on this comparison, we observed that the human eye is better in capturing objects in extreme brightness and contrast conditions than the machine. We think that this might be due to some extra information (e.g. experience) involved in human recognition.