Why is single image processing not performing as well as the mobile app?

Mobile app (barcode picker api) uses a video stream to scan. Scanning from a single image will never perform as well as scanning from a video stream, as the probability of getting a good picture of a code is much lower with just one image. In addition, barcode scanner can't control the camera based on previous frames if a single image is used. Real time adjustment of autofocus and exposure which is optimised for barcode scanning make a significant difference.