Illumination Normalization Using Weighted Gradient Integral Images

Abstract

Abstract-Captured images arise through interaction between objects of interest and illuminating light sources. If the latter are unevenly distributed, or are too strong or too weak, the image can have low contrast either locally or globally, impeding its interpretation and reducing its usefulness. In practice, control of illumination conditions is challenging, and not always possible. Thus, we propose a novel method to post-process captured images to reduce the effects of the illumination. We employ the Sobel operator to estimate gradients in the image, then use these gradients as weights in an averaging operation. To accelerate the operation, and remove certain unwanted features, integral images are iteratively calculated from the weighted images. This allows us to estimate the illumination component of the images, and in turn the reflectance component as their enhanced ones. A comparative study using a large number of images shows that the proposed algorithm outperforms several state of the art approaches.

Illumination compensation for facial feature point localization in a single 2D face image

“…Although all above-mentioned methods can partly eliminate uneven illumination in face image and provide barely satisfactory experimental data for subsequent recognition task, they all have the same deficiency that they do not perform well for the shadow around nose. Figure 1 shows some representative examples from references [1][2][3][4]. We can see that the shadow around nose, which has negative effect on our further image analysis, is obvious.…”

“…Among them, the most representative factor is uneven illumination. In order to address the problems due to uneven illumination, researchers have done some relevant works [1][2][3][4]. Vu and Caplier [1] proposed a new approach of illumination normalization based on Retina modeling by combining two adaptive nonlinear functions and a difference of Gaussians filter.…”

“…(d) Result from ref. [4] The rest of our paper is organized as follows. In Section 2, the framework and details of our proposed illumination compensation method for a single 2D face image are given.…”

“…Liu et al. [4] employed Sobel operator to estimate gradients in face images, and then utilized these gradients as weights in an averaging operation.…”

Multi-Scale Spatially Weighted Local Histograms in O(1)

“…Integral histogram provides an optimum and complete solution for the histogram-based search problem. Since then many novel approaches have been presented based on integral histogram to accelerate the performance of image processing tasks and incorporate the spatial information includ- ing filtering [8,9,10,11], classification and recognition [3,12], and detection and tracking [13,14]. Despite all different techniques that have been proposed to adaptively weight the contribution of pixels when computing local histograms by considering their distance from center pixel, the problem of how accurately extract the spatially weighted histogram of any arbitrary region within an image in constant time using integral histogram is still unsolved.…”