Study Of Image Processing Algorithm Based On Interest Area

With modern scientific further deepen research; the salient region in image has beenencouraged to research. The visual attention and related technology are used to quickly detectand extract the salient region in image, which has an urgent need in the military, the production,the life, and others fields, while it is also a key issues of computer vision, artificialintelligence, video communication and image processing, and other fields.The visual attention is an important psychological adjustment mechanism in the humaninformation processing. It is also a conscious activity of selecting the useful information from alarge amount of information, which is the efficiency and reliability of the guarantee in theprocess of human visual perception. The visual attention model based on the visual attention,and combined with the computer vision, which will form a visual attention system structure ofthe visual features. It can not only help to explore the vision cognitive laws, but also to solvethe problem on selecting the salient region, as well as to have an important significance ofimproving the efficiency of the image processing. In addition, it has important application valuein the image extraction and image scale fields.This paper will deeply research the image salient region, and based on the improve visualattention model, combined with the image processing algorithm, such as the image extractionalgorithm and the image scaling algorithm, will research the image salient region.Image salient region extraction is a research hot point in the field of the image processing.The salient region extraction technology will help instruct the image processing algorithm suchas the image scaling, the image mosaicing, can also help observer understand the image. For theextraction inaccurate problem of the image salient region, this paper has studied the imageextraction based on the visual attention model, and proposed an image salient region extractionalgorithm based on visual feature. Image salient region extraction algorithm based on visualfeature mainly extract the image region of visual feature obvious, and according to the differentimage, choosing different visual feature.Image scaling is one of the important parts in the image processing algorithm, which alsoplay an important role in the digital video processing field. Traditional methods often directlyscale image size in order to adapt to the target screen, but this method will lead to the imagesalient region fuzzy or distorted, and even to lose background information in image scalingprocess. To the above problems, this paper has studied the image scaling based on the salientregion, and proposed an image scaling algorithm based on salient region. Image scalingalgorithm based on salient region mainly narrow or enlarged the salient region in image. Thesalient region is the focus of research of this paper image scaling, is also the innovation points of this paper. The other region in image is not the focus of research, so may conduct the simplescaling processing.Based on the above algorithms, this paper has established an image processing algorithmexperimental platform based on the salient region, and has realized the above methods. Theexperimental results show:The improved visual attention model can not only accurately detect the salient region inthe images, but also provide the basis for the image salient region extraction and image scalingtwo kinds of algorithm in this paper.The image salient region extraction algorithm based on visual feature has not onlyextracted the image salient region and guaranteed the integrality of the salient region content,showing good visual effect, improving the quality of the image extraction.The image scaling algorithm based on salient region can not only achieve image resizing,and ensure the image no distortion and the integrity of salient region content, show good visualeffect, and improve the quality of image resizing.Finally, the whole work of this paper and the future research are concluded and discussed.