FCTH: FUZZY COLOR AND TEXTURE HISTOGRAM A LOW LEVEL FEATURE FOR ACCURATE IMAGE RETRIEVAL Savvas A. Chatzichristofis and Yiannis S. Boutalis Department of Electrical & Computer Engineering Democritus University of Thrace 12, Vas. Sofias, 67100 - Xanthi, Greece schatzic@ee.duth.gr, ybout@ee.duth.gr Abstract This paper deals with the extraction of a new low level feature that combines, in one histogram, color and texture information. This feature is named FCTH - Fuzzy Color and Texture Histogram - and results from the combination of 3 fuzzy systems. FCTH size is lim- ited to 72 bytes per image, rendering this descriptor suitable for use in large image databases. The pro- posed feature is appropriate for accurately retrieving images even in distortion cases such as deformations, noise and smoothing. It is tested on a large number of images selected from proprietary image databases or randomly retrieved from popular search engines. To evaluate the performance of the proposed feature, the averaged normalized modified retrieval rank was used. An online demo that implements the proposed fea- ture in an image retrieval system is available at: http://orpheus.ee.duth.gr/image_retrieval. 1. Introduction Content based image retrieval, known as CBIR, un- dertakes the extraction of several features from images, which, consequently, are used for the retrieval proce- dure. The visual content of the images is mapped into a new space, named the feature space. The features have to be discriminative and sufficient for the description of the objects. Basically, the key to attain a successful retrieval system is to choose the right descriptors that represent the images as “strong” and unique as possi- ble. Regarding their type, CBIR systems can be classi- fied in systems that use color information, those that use texture information and finally in systems that use shape information. It is very difficult to achieve satis- factory retrieval results by using only one of these fea- ture categories. Most of the so far proposed retrieval techniques adopt methods in which more than one fea- ture types are involved. For example, color and texture features are used in the QBIC [1], SIMPLIcity [2] and MIRROR [3] image retrieval systems. A question however that emerged the past years is how these features will become more compact. The characterization of an image with a high dimensional vector may have very good retrieval scores but it sig- nificantly delays the retrieval procedure and increases the storage requirements. The descriptors that were proposed by MPEG-7 [4] [5], for indexing and retrieval, maintain a balance be- tween the size of the feature and the quality of the re- sults. These descriptors appear to be in place to de- scribe well enough the visual content of the image. This paper proposes a new low level descriptor that includes in one quantized histogram color and texture information. This feature (FCTH) results from the combination of 3 fuzzy units. Initially the image is segmented in a preset number of blocks. Each block passes successively from all the fuzzy units. In the first unit, a set of fuzzy rules undertake the ex- traction of a Fuzzy-Linking histogram [6]. This histo- gram stems from the HSV color space. Twenty rules are applied in a three-input fuzzy system in order to generate eventually a 10-bin histogram. Each bin cor- responds to a preset color. As second unit, this paper proposes a two-input fuzzy system, in order to expand the 10-bins histogram into 24-bins histogram, importing thus information re- lated to the hue of each color that is presented. The process is described in section 2. Next, in the third unit, each image block is trans- formed with Haar Wavelet transform and a set of tex- ture elements are exported. These elements are used as inputs in a third fuzzy system which converts the 24- bins histogram in a 192-bins histogram, importing tex- ture information in the proposed feature. In this unit, eight rules are applied in a three-input fuzzy system and the process is described in section 3 and 4. With the use of the Gustafson Kessel [7] fuzzy clas- sifier, 8 regions are shaped, which are then used in or- Ninth International Workshop on Image Analysis for Multimedia Interactive Services 978-0-7695-3130-4/08 $25.00 © 2008 IEEE DOI 10.1109/WIAMIS.2008.24 191