3D image generation from textured digital images using improved linear algorithm based on depth map estimation and resolution enhancement

Abstract

Multimedia applications have gone through wide advancements in the recent years with the development of portable digital devices. With the advancements in multimedia application the demand for 3D technology has also increased. The most important issue in two dimension to three dimension image conversion is the poor image quality and its increased time complexity. The visual perception of the image can be improved by using pre-processing techniques on the noisy, blurred texture image with low resolution to remove spectral and spatial problems. The paper, explains the enhancement schemes using Discrete Wavelets and Stationery Wavelet Transforms with various interpolation techniques used as preprocessing technique. For the 3D conversion an Improved Simple Linear Iterative Clustering (ISLIC) method with Statistical Region Merging (SRM) is proposed. To retain the image quality Gaussian smoothing is performed with color uniformity principle. Here in the proposed work to reduce the time complexity a Depth Image Based Rendering (DIBR) method is applied to construct a 3D image from the given input. The performance analysis of this work has been compared with that of the existing methodologies and found to be more efficient.