MIQM: Multicamera Image Quality Measures

Mr. Rahul P. Bembade, Prof. Pankaj R. Chandre

Abstract With the rising demand of multiview application, quality assessment of multicamera images and videos is becoming crucial to the development of these applications. Image quality is depending upon several factors, which may be camera configuration, stability of camera during taking photo, number of cameras, and the calibration process etc. While numerous subjective and objective quality measurement methods have been projected in the literature for all images and all videos from single cameras, no comparable effort has been dedicated to the quality assessment of multicamera images. With the intention of develop an objective metric specially designed for multi camera system, we recognized and quantify two types of visual distortion in multicamera image: one is photometric distortions and another is geometric distortions. The comparative distortion between individual camera scenes is a major factor in determining the overall perceived quality. The distortions can be translated into components like contrast, luminance, edge based structure and spatial motion. We suggest three different indices that can compute these components. We give examples to show the correlation among these components and the corresponding indices. Multicamera image quality measure (MIQM) is calculated by combining three indices which are luminance and contrast index, spatial motion index and edge based structure index. The result and comparison with the other measures, like peak signal-to noise ratio (PSNR), mean structural similarity (SSIM), and visual information fidelity (VIF) prove that MIQM surpass other measure in capturing the perceptual fidelity of multicamera images. And in last, the results against subjective assessment are verified. Keywords: Fidelity measures, image quality assessments, multicamera arrays, multiview imaging, perceptual quality

Title:	MIQM: Multicamera Image Quality Measures
Author Name:	Mr. Rahul P. Bembade, Prof. Pankaj R. Chandre
Abstract:	Abstract With the rising demand of multiview application, quality assessment of multicamera images and videos is becoming crucial to the development of these applications. Image quality is depending upon several factors, which may be camera configuration, stability of camera during taking photo, number of cameras, and the calibration process etc. While numerous subjective and objective quality measurement methods have been projected in the literature for all images and all videos from single cameras, no comparable effort has been dedicated to the quality assessment of multicamera images. With the intention of develop an objective metric specially designed for multi camera system, we recognized and quantify two types of visual distortion in multicamera image: one is photometric distortions and another is geometric distortions. The comparative distortion between individual camera scenes is a major factor in determining the overall perceived quality. The distortions can be translated into components like contrast, luminance, edge based structure and spatial motion. We suggest three different indices that can compute these components. We give examples to show the correlation among these components and the corresponding indices. Multicamera image quality measure (MIQM) is calculated by combining three indices which are luminance and contrast index, spatial motion index and edge based structure index. The result and comparison with the other measures, like peak signal-to noise ratio (PSNR), mean structural similarity (SSIM), and visual information fidelity (VIF) prove that MIQM surpass other measure in capturing the perceptual fidelity of multicamera images. And in last, the results against subjective assessment are verified. Keywords: Fidelity measures, image quality assessments, multicamera arrays, multiview imaging, perceptual quality
Cite this article:	Mr. Rahul P. Bembade, Prof. Pankaj R. Chandre , " MIQM: Multicamera Image Quality Measures " , International Journal of Emerging Trends & Technology in Computer Science (IJETTCS), Volume 3, Issue 4, July - August 2014 , pp. 244-251 , ISSN 2278-6856.
Full Text [PDF] Home