The Automatic Recognition And Research Of Analytical Methods To The Glaucoma Images

Glaucoma is one of the most common diseases that cause blindness, which is characterized by elevated IOP, atrophy of the optic nerve and defect of visual field. Because of the stealthness and incrementality of the glaucoma, especially Primary Open Angle Glaucoma, once detecting impaired vision and seeing a doctor, it's a late stage, the defect of visual field is serious and can't be recovered. In view of the above, earlier discovery of the glaucoma should be emphasized. At present, the early detection of glaucoma is still a difficulty, staying at the state of resorting to experiences, which has a big deviation. In order to realize the automation of recognition, it's necessary to develop a kind of feasible recognition method.This article has first introduced the actuality of glaucoma recognition and pointed out its necessity, then taking computer image processing as main technique measure, has studied the methods "of automatic recognition, integrated image processing, mathematical morphology and neural network.Taking the map of glaucoma images as example, and main research work is as follows:1. Collected 54 sheets of glaucoma images from the map, including 8 sheets of eye-ground images of normal eyes, 8 sheets of eye-ground images of early glaucoma eyes and 38 sheets of eye-ground images of late-stage glaucoma eyes.2. Studied the preprocessing methods of images, including de-noising methods enhancement processing.3. Primarily extracted 49 characteristic values from color, geometry shape and monochrome statistic on processed images, the picked out 32 typical characteristic values resorting to PCA (principal component analysis).4. Took advantage of neural network to set up classifier and carry on classification and identification to the images based on MATLAB.5. Taking the MATLAB GUI as development tool, compiled "processing and recognition system of glaucoma images" , realized the automation recognition to the glaucoma images.This article realized the recognition to the early glaucoma images, which is fast, general-purpose and accurate. It is doubtless that this new method is an enormous improvement in the recognition method based on manual work, so it has vasty potential application in the field of early recognition of glaucoma.