Recognition of blurred plate numbers using a novel algorithm based on Hopfield neural network

In this paper, we propose a novel algorithm based on the Hopfield neural network (HNN) and apply it to the recognition of blurred plate numbers. For a blurred plate number with m characters, the proposed algorithm takes competitions in pairs for each given pattern of m groups by using the HNN. The won numbers of given patterns of m groups can be obtained. Utilizing these won numbers, the competition score of every plate number can be calculated. According to these competition scores, we can sort the plate numbers and select the former sorted plate numbers as candidates. This result is very helpful for the criminal investigation organization to reduce the investigation range and is available to investigate these candidates one by one