3D Image Retargeting

Miss Umaima Bohari, Prof. S. M. Kamalapur

ABSTRACT Advancement in 3D imaging technology has increased the capture and display of 3D images. 3D images can be viewed over many devices, ranging from high-resolution computer monitors, cinema screens to low-resolution mobile devices. These images are also processed by many computer vision algorithms. Stereoscopic images often have to experience changes in size as well as aspect ratio in order to adapt to different display screens. Displaying and printing documents with embedded images often involve resizing of the images to fulfill the general layout. Image resizing by cropping, scaling or stretching often do not generate satisfactory and pleasing results, distorting image structure and 3D depth. This work focuses on reviewing and categorizing algorithms for content aware 3D image retargeting which means resizing an image while taking its important content into concern to preserve its salient regions and minimizes the visual and depth distortion. Also modified stereo seam carving method is proposed for retargeting 3D images into required aspect ratio. There are many challenges in 3D image retargeting as compared to 2D because focus is to maintain stereo consistency as well as relevant information in the image. The proposed technique work by computing energy function for every pixel or group of pixels in order to find their importance and then apply an operator that retarget the image taking into consideration the importance map, depth map and additional constraints. Keywords: Depth map, Depth distortion, Stereoscopic 3D images, Salient regions

Title:	3D Image Retargeting
Author Name:	Miss Umaima Bohari, Prof. S. M. Kamalapur
Abstract:	ABSTRACT Advancement in 3D imaging technology has increased the capture and display of 3D images. 3D images can be viewed over many devices, ranging from high-resolution computer monitors, cinema screens to low-resolution mobile devices. These images are also processed by many computer vision algorithms. Stereoscopic images often have to experience changes in size as well as aspect ratio in order to adapt to different display screens. Displaying and printing documents with embedded images often involve resizing of the images to fulfill the general layout. Image resizing by cropping, scaling or stretching often do not generate satisfactory and pleasing results, distorting image structure and 3D depth. This work focuses on reviewing and categorizing algorithms for content aware 3D image retargeting which means resizing an image while taking its important content into concern to preserve its salient regions and minimizes the visual and depth distortion. Also modified stereo seam carving method is proposed for retargeting 3D images into required aspect ratio. There are many challenges in 3D image retargeting as compared to 2D because focus is to maintain stereo consistency as well as relevant information in the image. The proposed technique work by computing energy function for every pixel or group of pixels in order to find their importance and then apply an operator that retarget the image taking into consideration the importance map, depth map and additional constraints. Keywords: Depth map, Depth distortion, Stereoscopic 3D images, Salient regions
Cite this article:	Miss Umaima Bohari, Prof. S. M. Kamalapur , " 3D Image Retargeting " , International Journal of Application or Innovation in Engineering & Management (IJAIEM), Volume 3, Issue 10, October 2014 , pp. 238-242 , ISSN 2319 - 4847.
Full Text [PDF] Home