The Research Of Image Retrieval Method Based On Region Of Interest

With the rapid development of the Internet,the type and number of images is increasing.The image has become an important channel of access to information in people's work and life.Comparing to text messages,images convey the information more intuitive,rich,scene reduction more accurate,so the image information obtained from the more favored by the people.Therefore,as an image to obtain accuratel channel,image retrieva has been rapid development and become one of the important research areas of information retrieval.The traditional image retrieval method is a natural transition from the text retrieval method,we need to treat retrieved images manually tag.Such an approach in the case of retrieving the image can be used less.As the image geometric order of magnitude increase in the number of artificial way mark takes enormous human and material resources,it can not work.Content based image retrieval emerged,which accord to the expression of the realization of the contents of the image example inquiry.Currently content based image retrieval methods mostly use low-level features of the image describe the image,ignoring the high-level semantic features of the image,resulting in the "semantic gap" brings retrieval results less than expected.How to narrow the "semantic gap",depicts a user to retrieve exact needs,improve the retrieval effect,which becomes one of the current hot topics.The image retrieval methods existing usually merge single or global features to realize the retrieval.Image retrieval method based on the features fusion and adapt to complementary of region of interest and global image.So,we combine with the current machine learning methods in image processing to propose an image retrieval method based on region of interest,which can improve precision rate and recall rate of image retrieval,reflecting the query intent.This paper focuses on the region of interest for image retrieval,and it consists of the low-level and high-level image feature extraction,similarity calculation method,image retrieval framework etc.The main innovative works of this paper can be summarized as follows: Detailed study of multi-feature integrated search feature to solve the multi-image retrieval feature selection,multi-feature similarity measure integration issues.To make up for the use of image retrieval using a single feature can not be effectively retrieved to meet the demands of image retrieval.Achieve perceptual algorithms to extract the image region of interest based on visual,and for image retrieval.Based on eye-tracking data presented eye-tracking device to capture images of the user observation,the image region of interest extraction algorithm based on eye-tracking data,and the combination and image retrieval.And experiment show the effectiveness of the algorithm.