An Image Based on SVM Classification Technique in Image Retrieval

On the frontier of Image Processing, researchers are encountering the challenge of effectively retrieving and using the information contained in the image. As per the prevailing research after the feature extraction of the relevant properties of a high-level image, the resulting image does not add too many features, When operating directly on the image, because of the high Witte sexual performance data are relatively poor, resulting in the traditional classification method does not apply. So this paper uses support vector machine (SVM) image classification techniques which can overcome this defect. This paper makes the use of Dense SIFT algorithm to obtain image feature and then build Bag of words model. Subsequently establishing training dictionary database and finally, the test set of images SVM classification test. Experimental results show that the use of SVM classification accuracy of image retrieval technology enables greatly increased.