Video dynamic range compression of portrait images by simulated diffuse scene illumination

Abstract

The mismatch between dynamic ranges of intensities in nonuniformly illuminated portrait scenes and of intensities reproducible on common display devices is addressed. In particular, such mismatches appear quite often in videophone and videoconferencing situations. Displayed portrait images with no details visible in some facial regions result from them, and lead in turn to a reduced impression of telepresence. A novel image processing approach for dynamic range compression of portrait images is presented in this paper. Unlike conventional methods, it provides compressed images with a most natural appearance. The basic idea is to simulate a more uniform scene illumination in the video signals of portrait images. This requires knowledge about the scene's reflectance function. Based on scale-space filtering, an edge discrimination technique is developed that distinguishes intensity edges caused by reflectance from edges caused by illumination at smoothly curved surfaces. This facilitates an estimation of the reflectance component from the intensity distribution of portrait images using an extension of the classical lightness computation approach

Lightness determination at curved surfaces with applications to dynamic range compression and model-based coding of facial images

“…However, they provide dynamic range compressed portrait images that lack naturalism [4]. 11 This would again counteract the effort of video communications to achieve telepresence.…”

“…In (3), is a vector field, and thus is a vector field, too, wherein small gradients of have been suppressed, but large ones have been retained. Subsequent reintegration of leads to the recovery of log-reflectance up to an unknown additive constant , as follows: (4) This is the lightness we are after. One obtains the scene's relative reflectance function by exponentiation of (4).…”

“…In general, about 100 iterations were required to reach a stable solution. [4]. An efficient multigrid relaxation technique may be used for this task [29].…”

“…Both components combine as a product to form 4 (1) is assumed to be the sum of two components [6]: Directional irradiance produced by light originating from the light source(s) with no intermediate reflections and ambient 3 This leaves aside light emitting surfaces that do not appear in most real world scenes. 4 A factor of proportionality depending, among others, on the optical arrangement is neglected. or indirect irradiance caused by light reflected already from surfaces in the environment.…”

Quantitative analysis of infrared contrast enhancement algorithms

“…In essence, dynamic range compression is a pixel operator, defining the value of an arbitrarily located pixel in a new image by using the value of the corresponding pixel in the original image. Dynamic range compression has a number of applications in fields such as video telephony [1], radiology [6,11], and high dynamic range photography [7,9]. As such, research has been performed with the aim of providing a dynamic range compression algorithm that suits an application area's needs, such as enhanced image quality, heightened information availability, or real-time processing capability.…”