A Novel Approach for Enhancing QoS for Visualization of High Dynamic Range Imaging

High dynamic range (HDR) imaging has gained a lot of interest as a technology to mimic human visual features as a result of advancements in display and camera technology. Prior schemes have explored the issues of low accuracy of local image data. When lighting is poor, then the HDR images appear vague that lead to poor image analysis. Similar to it, the image distortion could deprive the image information. Hence, a novel image fusion framework is designed by combining the dense SIFT descriptor and improved guided filter. The guided filter has improvised the tone mapping process so as to preserve the edges of an image. Initially, the dense SIFT descriptor extracts the local image data from the guided images. In the case of dynamic scenes, the formation of histogram equalization, median filtering and Gaussian filtering are employed to estimate the color dissimilarity feature. The initial weights used to hybrid the Gaussian and median filters are done by image data such as contrast, brightness and the dissimilarity features. Henceforth, the adoption of improved guided filters is employed to eliminate the noise and the discontinuity of the initial weights estimation. Experimental results of the proposed techniques have shown the better results with respect to different evaluation metrics.