Prof. Alan Bovik, professor in the Department of Electrical and Computer Engineering at University of Texas at Austin’s Cockrell School of Engineering along with Lark Kwon Choi, doctoral student in ECE have developed a no-reference perceptual fog density prediction model and a perceptual image defogging algorithm that are based on natural scene statistics (NSS) and fog aware statistical features.
Together, the perceptual fog density prediction model and the perceptual image defogging algorithm can help quantify the level of fogginess in an image of a scene as well as help in enhancing visibility by defogging the image. In order to do this, the model uses a previously stored fog aware collection of statistical data, which is derived from a corpus of foggy and fog-free images containing explicit details about its characteristics such as low contrast, faint color and shifted intensity for better detectability and defogging. The predictions are based on algorithmic data that is derived by comparing foggy and fog-free images to measure statistical irregularities.
“We are incorporating elements of visual perception concepts from neuroscience, which can predict the visibility of any foggy scene without any extra information. This model uses previously stored insights about fogginess or the lack of it as perceived by the visual brain”, said Prof. Bovik.
Prof. Bovik is the Ernest J. Cockrell Endowed Chair in Engineering at the University of Texas at Austin. He is also a professor in the Institute for Neurosciences, and Director of the Laboratory for Image and Video Engineering (LIVE) at UT. He also believes that the Perceptual Fog Density Assessment and Image Defogging technology could be useful for firefighting rescue operations to help find people and guide fire fighters.
“In the future, we are also hoping to integrate this technology with biometric applications such as facial recognition, or program it to detect certain objects, which will then make it useful for surveillance purposes”, added Lark.
Lark is currently pursuing his Ph.D. as a member of the Laboratory for Image and Video Engineering (LIVE) at The University of Texas at Austin under Prof. Bovik’s mentorship. His research interests include image and video quality assessment, spatial and temporal visual masking, motion perception, and perceptual image and video enhancement.
The paper titled, ‘Referenceless Prediction of Perceptual Fog Density and Perceptual Image Defogging, which will explain the working and application of this technology in greater detail will soon be published in the IEEE Transaction on Image Processing 2015.