Research Of Shoeprint Images Retrieval Algorithm And System Design

Shoeprint image are often left at crime scenes and can be used as scientific evidence in forensic science. Recent developments in forensic science have resulted in large numbers of scene of crime images being collected for recording and analysis, so rapid and effective searching for desired images from large-scale shoeprint image databases becomes an important and challenging research topic.In this thesis, the exploratory research work has been done around the feature extraction, which include texture, shape. For the widespread incomplete of shoeprint image, the shoeprint image is divided into several parts before extracting the texture features and evaluate the effectiveness of each part. Only when the partition is a valid part, we extract texture features and participate in retrieval. On texture feature extraction aspects, two retrieval algorithms are presented. One is improved texture feature retrieval algorithm based on Log-Gabor filter cluster, another is projection Fourier amplitude spectrum texture feature retrieval algorithm based on auto-registration of Fourier Transform magnitude spectra. On shape feature aspects, a kind of shoeprint image retrieval algorithm based on the match of shape area is proposed.At the same time, a kind of shoeprint image classification algorithm based on shape and statistical feature is presented. In this algorithm, shoeprint image is divided into the type of suitable for extraction texture features and the type of suitable for extraction shape features, so it can choose a suitable method in features extraction.A system of shoeprint image retrieval is designed and realized on the basis of it. The system use Access2003 as a database and programming implementation in VC6.0, which can retrieval image using the methods of this paper.