This study introduced an automatic authentication technique for checking the genuineness of a vehicle. The rough set-based technique was used to handle the uncertainty arisen from artifacts in the acquired images impr...This study introduced an automatic authentication technique for checking the genuineness of a vehicle. The rough set-based technique was used to handle the uncertainty arisen from artifacts in the acquired images imprinted on a vehicle. However, it has been proved to be NP-hard to find all reductions and the minimal reduction, and generally different heuristic algorithms were used to find a set of reductions and the Ganssian distribution was used to describe the uncertainty to achieve the minimal reduction. On the basis of inductive logic programming, the technique can distinguish between two similar images, as is superior to the conventional pattern-recognition technique being merely capable of classifier. Furthermore, it can avoid some failures of the technique based on the correlation coefficient to authenticate binary image. The experiments show an accuracy rate close to 93. 2%.展开更多
基金Sponsored by the National High Technology Research and Development Program of China (Grant No. 2003AA1Z2610)
文摘This study introduced an automatic authentication technique for checking the genuineness of a vehicle. The rough set-based technique was used to handle the uncertainty arisen from artifacts in the acquired images imprinted on a vehicle. However, it has been proved to be NP-hard to find all reductions and the minimal reduction, and generally different heuristic algorithms were used to find a set of reductions and the Ganssian distribution was used to describe the uncertainty to achieve the minimal reduction. On the basis of inductive logic programming, the technique can distinguish between two similar images, as is superior to the conventional pattern-recognition technique being merely capable of classifier. Furthermore, it can avoid some failures of the technique based on the correlation coefficient to authenticate binary image. The experiments show an accuracy rate close to 93. 2%.
