Immersive imaging

• Main Author: Sow, Hendy Sutomo.
• Other Authors: Sabu Emmanuel
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/38615
• Institution: Nanyang Technological University

Immersive imaging is a project that lets users to experience a virtual reality effect while doing a zooming of an image. As this program provides a smooth transition from one layer of the image to the other, users will experience a sense of traversing inside the image. This is achieved by interpolating image of the first layer and the second layer to obtain image in the intermediate. Images taken with camera are required as input. They must be images of an object with variation of focal length setting of the lens. The purpose is to provide images with different zooming level of a single object. Before interpolation is applied, mapping of the pixels of first layer image (original image) to second layer image (zoomed image) need to be calculated. As observed in zoomed image, objects that are located near the border of original image have been omitted due to the zooming effect. If interpolation is applied directly, output image will not contain the correct pixel values. Consequently, it will generate a random unrecognized image. The mapping method that is applied is called critical-points filters [1] [2]. The first step of the mapping is to extract the critical points (maxima, minima and saddle points) from the source and destination image. After the extraction, there will be sub-images (images with lower resolution) that contain each of the critical points. From these sub-images, mapping is calculated based on an energy function. The correct mapping is the mapping with the lowest energy.