Seyed Muhammad Hossein Mousavi - Department of Computer Engineering Bu Ali Sina University Hamadan, Iran, ۶۵۱۴۱

Detection and recognition have higher accuracy using learning-based methods but not the fastest. Template matching design is faster than leaning based design. Especially when there are no enough leaning data, there are no options except using matching methods. This paper tries to introduce a method for finger print recognition using few images per subject based on template matching techniques out of object’s surface. Here there is just one sample as target and there might be few or even just one sample as input to match. The point here is the speed and mobile capability of the system but the most important part of this research is to using magnifier which is added to the capturing device which is an ordinary phone here. LG-K10 2017 smart phone is used in this research as main capturing device. Also, system is android based which with adding a magnifier could be a proper cheap field experiment tools in surveillance and security applications. To have better fingerprint recognition out of object’s surface, proper preprocessing techniques helps to have higher accuracy which this paper employed this fact very well. Also, this research does not use any chemical material like powder to aid the recognition process which has cost and time advantages. experiments show that matching using magnifier and proper preprocessing techniques have 25% improvement versus normal capturing.

Template matching; Finger print recognition;object’s surface; Smart phone; Magnifier; Surveillance and security