The web app can detect books in a book shelf image by segregating and digitizing book spines. The app uses canny edge detection and hough line transformation to identify book spines. Then crops the image on the basis of these book spines to digitize the book information through google vision API.
Here is the demo: