Medical Image Retrieval Based On PCNN

In the medical area, with the widely use of X-ray, CT, MRI and other medical imaging equipment, medical image processing technology has become a basic technology in modern medicine. In clinical diagnosis, medical teaching and research activities, people need large amounts of medical images, resulting a huge demand of a effective way to manage and retrieval the medical image data. With the demand of medical image query growing, how to manage the large mount medical image database qualitatively has become an urgent problem to be solved.Content-based image retrieval, CBIR for short, is a branch of large amount digital image content retrieval in the field of computer vision. The CBIR technology analyze and classified the image in database, extract the feature of the image by image model and stored them in the image feature database, while indexing the feature to improve search efficiency. The user can set query conditions through the user interface, they can use one or several features, then the system use the similarity matching algorithm to calculate the similarity between the key features and features in the image database,at last the system feedback the query result order by the similarity to the user. The CBIR technology solve the problem of large number of images retrieval, and it’s a meaningful work to introduce the content-based retrieval technology into the field of medical research.This paper focus on the content based image retrieval, studied the content based medical image retrieval and the use of pulse coupled neural network in CBMIR.This paper propose a medical image retrieval algorithm based on PCNN. This algorithm use the PCNN to process the medical image, and extract the Euler number of each image int the image sequence generate by the PCNN. The experimental result of1200medical images shows, this algorithm is better than the common feature extraction method based PCNN.This paper improved the neuron model based the research of the proposed algorithm, reduce the number of the parameters, simplified the structure. The experimental result shows the modified neuron model is basically the same as the original PCNN model, but it’s has more simple parameter settings and faster calculate speed.This paper designs and implements a content-based medical image retrieval prototype system, with this system, the algorithm can be easily verified. The proposed algorithm has been verified by the prototype system and1200medical images.