<title>Image retrieval and reversible illumination normalization</title>

Abstract

We propose a novel approach to retrieve similar images from image databases that works in the presence of significant illumination variations. The most common method to compensate for illumination changes is to perform color normalization. The existing approaches to color normalization tend to destroy image content in that they map distinct color values to identical color values in the transformed color space. From the mathematical point of view, the normalization transformation is not reversible.In this paper we propose to use a reversible illumination normalization transformation. Thus, we are able to compensate for illumination changes without any reduction of content information. Since natural illumination changes affect different parts of images in different amounts, we apply our transformation locally to sub-images. Basic idea is to divide an image into sub-images, normalize each one separately, and then project it to an n-dimensional reduced space using principal component analysis. This process yields a normalized texture representation as a set of n-vectors. Finding similar images is now reduced to computing distances between sets of n-vectors. Results were compared with a leading image retrieval system.

Detection of emotional faces: Salient physical features guide effective visual search.

“…We computed basic image statistics, such as mean luminance, contrast density (root-mean-square contrast), and global energy (see Kirchner & Thorpe, 2006) with Matlab 7.0 (The Mathworks, Natick, MA). In addition, we computed color and texture similarity with local pixel-by-pixel principal component analysis with reversible illumination normalization (see Latecki, Rajagopal, & Gross, 2005).…”

Visual Search of Emotional Faces

“…To this end, we assessed (a) mean luminance and (b) contrast density (RMS contrast, Bex & Makous, 2002) of each face stimulus by means of Adobe Photoshop. In addition, we assessed (c) color and (d) texture similarity each target (emotional) face and the corresponding context (neutral) face, by implementing a local pixel-by-pixel principal component analysis (PCA) with reversible illumination normalization (see Latecki, Rajagopal, & Gross, 2005). One-way ANOVAs (type of emotional expression) yielded no significant differences in luminance ( p = .61) or RMS contrast ( p = .35) among the emotional faces.…”

Building Roof Segmentation from Aerial Images Using a Lineand Region-Based Watershed Segmentation Technique

“…For this, 24 colorimetric invariants in the literature have been tested: greyworld normalization (called greyworld in Figures 13 and 14 ) [ 45 ], RGB-rang [ 46 ], affine normalization (called affine in Figures 13 and 14 ) [ 47 ], intensity normalization (called chromaticity in Figures 13 and 14 ) [ 46 ], comprehensive color normalization (called comprehensive in Figures 13 and 14 ) [ 45 ], c1c2c3 [ 40 , 41 ], m1m2m3 [ 41 ], l1l2l3 [ 40 , 41 ], l4l5l6 [ 48 ], A1A2A3 [ 43 ], c4c5c6 [ 48 ], HSL, MaxRGB [ 46 ], CrCgCb [ 43 ], color constant color indexing (called CCCI in Figures 13 and 14 ) [ 46 ], m4m5m6 [ 43 ], standard L2 (called L2 in Figures 13 and 14 ) [ 43 ], maximum-intensity normalization (called Mintensity in Figures 13 and 14 ) [ 49 ], reduced coordinates [ 50 ], CrCb [ 40 , 43 ], opposite colors (o1o2) [ 40 , 43 ], saturation (S) [ 41 ], log-hue [ 46 ] and hue (H) [ 41 , 50 ]. Figure 3 illustrates the visual difference of some colorimetric invariants applied on the initial image.…”