Investigation Of Adaptive Zooming Method For Content-based Picture Size

Image resizing is a basic and important technology in Digital Image Processing. Linear resizing is a common method of image resizing. At present commonly used method is linear scaling method, just mechanical click the image size to zoom, therefore there exists the image distortion shortcomings. The Seam Carving technology, based on content, whichi is currently very effective solution can perception adaptive image scaling method, through the continuous removed or insert image of content in less effect, can change the image seam, realize the length of the image reconstruction, keep image the interested region, realizes image target removed, realize the size of the image magnification, and be able to shrink and keep images in zoom process.This paper analyzed and studied the image scaling method based on seam carving technology, through define the energy function, compute index matrix, give single and double coordinates direction transform ways, realizing "Calculate only one time, and use time after time" method, improve seam carving technology efficiency; put forward a new kind of significant regional detection and image segmentation method, this method gives significant area detection model, realized the significant regional segmentation, and give these regional different weights, by choosing proper threshold, using the method of image segmentation image remained important area, then put significant test results, and add weight to the energy function division, enhance energy function in images important area of protection; through the improvement of the original scaling algorithm, given the scale of the detection, significant regional segmentation algorithm, to avoid distortion of the contents of the user region of interest situations, to achieve the specified target region of interest removed and retained.Experimental results show that research can not only effectively scaling the image, but also successful completion of a user-specified region of interest to retain and to remove the specified target, based on Seam Carving proved natural image processing technology in the higher application value.