A novel approach for matching composite sketches to mugshot photos using the fusion of SIFT and SURF feature descriptor

Identification and apprehension of criminals by matching facial sketches with photographic faces is one of the major law enforcement applications of the modern world. Majority of the crime occur where there will not be any information available regarding the suspect. In such situation, forensic sketch artist who usually deal with the eyewitness of the crime or victim in order to draw the sketch that resembles criminal face according to the verbal description given by the eyewitness. These sketches are termed as `forensic sketches' and can be matched manually against gallery of criminal mug shot photos. By developing an automatic system for matching these sketches to the criminal database would reduce many person-hours. The paper presents a novel feature based approach which measures the similarity between sketches and mugshot photos. An efficient preprocessing technique is implemented on sketches and photos. Speed Up Robust Features (SURF) and Scale Invariant Feature Transform (SIFT) features are extracted and the resultant features are matched using nearest neighbor algorithm. Experiment are carried out using 40 composite sketches with 220 photos and it is observed that our two approach with image preprocessing given assuring results with good accuracy.