Abstract:
Person identification has really advanced with advancement in technology; whereby every feature of the body is reckon to be very important in identifying a person as their characteristics are usually unique to the bearer. The use of foot print which is one of the body features in person identification is inexhaustible, as much had not been done with it compared with other body features. Considering the various research works that has been carried out on footprint, it is certain that individuals can be traced to their prints, a major gap that research works have not really considered is the possibility of tracing a partial footprint to a person especially in case of crime scene or accident. Hence this work considered identifying individuals with partial footprint by designing and implementing a person identification system based on partial footprint. The method employed involves gathering footprints from volunteers and making the prints go through image processing techniques. The images were denoised and enhanced using Gaussian filters and K-means clustering, the features were extracted using principal component analysis. The extracted features from partial foot print were matched with those of whole footprint for similarities. The efficiency of this system was tested using time taken and standard performance metrics of Receiver Operating Characteristics, where 200 samples were used and 145 matched returning true positive. There is no false negatives and this means 98% specificity which can be interpreted as having perfect classification. Hence it is concluded that the system is reliable.
Keywords: Body features, foot print, identification, images, image processing technique.
