Does KONAN Cataloger support a face detection function?
Face images are automatically extracted by keyframe scanning. Various sizes of images with at least 20 x 20 pixels are produced. Because the face images contain timecodes identical to those of the keyframes, our users can locate and play the very scenes where a certain face appears. Although extracting face images is very complex and time-consuming, KONAN Cataloger makes it possible to catalogue videos much faster than realtime owing to Konan’s efficient algorithm. The accuracy of face detection is determined according to the amount of face-study data. The face-study data is an optional item for more accurate face detection.