A wireframe-detection based method for correcting credential photos

To correct credential photos like ID cards and bank cards for both saving the storage space and improving their quality, it usually needs to detect their corresponding border lines. Prevailing line-detection methods can be applied to detect the border lines,while they may also detect many other lines which are not border lines (denoted as non-border lines later) and lead to both lower efficiency and worse accuracy. This paper presents a wireframe-detection based method (WDBM) by combining with a line detection method based on the concept of “discrimination value”, for correcting credential photos like ID cards and bank cards. It can automatically discard many non-border lines and quickly find the correct border lines. Experimental results show that the WDBM can recognize part of lines that the prevailing line-detection methods do not recognize, and the WDBM can also achieve a better PSNR by refining the accuracy of the detected lines. In a word, both efficiency and accuracy of the WDBM are improved by comparing with prevailing line-detection methods.