ICTEDCT-CBIR: Integrating curvelet transform with enhanced dominant colors extraction and texture analysis for efficient content-based image retrieval

doi:10.1016/j.compeleceng.2012.05.010

Computers & Electrical Engineering

Volume 38, Issue 5, September 2012, Pages 1358-1376

https://doi.org/10.1016/j.compeleceng.2012.05.010 Get rights and content

Abstract

A novel Integrated Curvelet-based image retrieval scheme (ICTEDCT-CBIR) has been proposed, for the purpose of effectively retrieving more similar images from large digital image databases. The proposed model Integrates Curvelet Multiscale ridgelets with Region-based vector codebook Subband Clustering for enhanced dominant colors extraction and texture analysis. An important ingredient of the curvelet transform is to restore sparsity by reducing redundancy across scales. The discrete curvelet transform makes use of a dyadic sequence of scales, and a bank of filters with the property that the pass band filter is concentrated near the frequencies. An enhanced Region-based vector codebook Sub band Clustering (RBSC) has been proposed for effectively extract dominant colors from the color histogram of the transformed image sub-bands. An integrated matching scheme, based on most similar Highest Priority (MSHP) principle, is used to compare the query and target images. Experimental analysis has been carried out to verify the efficiency of the proposed ICTEDCT-CBIR model. Experimental results showed that the proposed approach has better retrieval performance. First, curvelets capture more accurate texture information. Second, as curvelets are tuned to different orientations, it captured more accurate directional features than wavelets. As the experimental results indicated, the proposed technique outperforms other retrieval schemes in terms of average precision with higher precision–recall crossover point values.

Graphical abstract

Highlights

► Integrates Curvelet Multiscale ridgelets with vector Subband Clustering. ► Combines spatial band pass filtering at different scales, and orientations. ► Structured dominant Colors with Higher average precision and recall values. ► More accurate directional features are captured, with observable performance gain. ► Retrieval performance is higher than that of Gabor and wavelet-based retrieval.

Introduction

Modern image search engines retrieve images based on their visual contents, commonly referred to as Content Based Image Retrieval (CBIR) systems [1], [2]. CBIR systems have found applications in various fields like fabric and fashion design, interior design as panoramic views, art galleries [3], [4], geographical information systems, remote sensing and management of earth resources, scientific database management, medical imaging, trademark and copyright database management, the military, law enforcement and criminal investigations, picture archiving and communication systems, retailing and image search on the Internet [4]. As the bandwidth availability increases to access the internet, users will be allowed to search for and browse through video and image databases located at remote sites [5]. For that reason we need fast retrieval of images from large databases which is a new point of research addressed nowadays. CBIR is an important alternative to traditional text-based image searching and can greatly enhance the accuracy of the information being returned. It aims to develop an efficient visual-content-based technique to search, browse and retrieve relevant images from large-scale digital image collections [6]. Image retrieval is the task of searching for images from an image database.

The CBIR process consists of calculating a feature vector that characterizes some image properties, and stored in the image feature database. The user provides a query image, and the CBIR system computes the feature vector for it, and then compares it with the particular image feature database images. The relevance comparison is done by using some distance measurement technique, and the minimum or permissible distances are the metrics for the matched or similar images. The features vector should be able enough to fully characterize image structural and spatial properties, which retrieve the similar images from the image database [2], [3], [4], [7].

In a CBIR system, the images are indexed by their visual contents as the features. These include the characteristics such as colour, texture, shape. When a query image is given, the features of the query image are extracted to match the features in the feature database by a pre-established algorithm, so that a group of similar images to the query image can be returned as the retrieval images. CBIR techniques use low-level features like texture, color, and shape to represent images and retrieves images relevant to the query image from the image database. Among those low level image features, texture feature has been shown very effective and subjective.

A number of texture features have been proposed in literature [8], [9], [10], including statistic methods and spectral methods. However, most of them are not able to accurately capture the edge information which is the most important texture feature in an image. Color is one of the most reliable visual features that are also easier to implement in image retrieval systems. Color is independent of image size and orientation, because, it is robust to background complication. Color histogram is the most common technique for extracting the color features of colored images [11], [12]. Color histogram tells the global distribution of colors in the images. It involves low computation cost and it is insensitive to small variations in the image structure. However, color histogram hold two major shortcomings. They are unable to fully accommodate the spatial information, and they are not unique and robust. Two dissimilar images with similar color distribution produce very similar histograms. Moreover, similar images of same point of view carrying different lighting conditions create dissimilar histograms. Many researchers suggested the use of color correlogram for avoiding inconsistencies involving the spatial information [11]. Multiresolution histograms [12], [13] are also suggested to ameliorate image retrieval process. Gaussian filtering may also be used for multiresolution decomposition of an image [13].

In this paper, a texture feature based on curvelet transform is proposed. The technique makes use of curvelet transform which represents the latest research result on multiresolution analysis [14], [15]. By combining the advantages of the two methods, image edge information is captured more accurately than conventional spectral methods such as wavelet and Gabor filters. Curvelet was originally proposed for image denoising and has shown promising performance. In this paper, we describe the theory and implementation of curvelet, then combine a Region-based vector codebook Subband Clustering (RBSC) for dominant color extraction with efficient curvelet-based sub-band texture extraction. Our proposed ICTEDCT-CBIR system is based on curvelet multiscale decomposition combined with a spatial band pass filtering operation to isolate different scales. Efficient color and texture feature extraction are proposed. The rest of the paper is organized as follows. A literature survey is introduced in Section 2. Section 3 presents an overview of digital curvelet transform.. In Section 4, the proposed ICTEDCT-CBIR image retrieval model will be introduced with illustration of model phases. Section 4.5 illustrates the similarity matching scheme. Experimental results and discussion are illustrated in Section 5 and, finally, conclusions are driven in Section 6.

Section snippets

Literature survey

A variety of techniques have been developed for extracting texture features. These techniques can be broadly classified into spatial methods and spectral methods. In spatial approach most techniques rely on computing values of what are known as low order statistics from query and stored images [8]. These methods compute texture features such as the degree of contrast, coarseness, directionality and regularity [8], [9]; or periodicity, directionality and randomness [4]. Alternative methods of

Digital curvelet transform

Curvelets are based on multiscale ridgelets combined with a spatial bandpass filtering operation to isolate different scales [24], [28], [30]. The Discrete Curvelet Transform decomposites an image into several levels of curvelet coefficients, each level has a number of sub-bands locate at different directions. Given an image function f(x, y), the continuous ridgelet transform is given as [27]: $R_{f} (a, b, θ) = \iint ψ_{a, b, θ} (x, y) f (x, y) dx dy$ where a > 0 is the scale, b ∈ R is the translation and θ ∈ [0, 2π] is the

The proposed ICTEDCT-CBIR image retrieval model

The proposed ICTEDCT-CBIR model will pass through the following phases as illustrated in Fig. 3. In the following sub-sections, each phase will be described.

A description of each phased of the proposed model will be illustrated in the following sections, including:

–
Transfer to HSV color space.
–
Apply multiscale curvelet transform to get different multiscale subbands.
–
The proposed Region-based Subband Clustering approach (RBSC). For each selected cluster i (i = 1, … , N_DCD), calculate the centroid, v_i.

Similarity matching

An integrated image matching procedure similar to the one used in [37], [38] is proposed. With the decomposition of the image, the number of sub-blocks remains same for all the images. In [6], a sub-block from query image is allowed to be matched to any subblock in the target image. However a sub-block may participate in the matching process only once. A bipartite graph of sub-blocks for the query image and the target image is built as shown in Fig. 6. The labeled edges of the bipartite graph

Experimental results

Experiments have been carried out to validate the efficiency of the proposed model. The experiments were carried out on a Core i3, 2.4 GHz processor with 4GB RAM using MATLAB 7.0. Comparisons with other methods have been conducted. The proposed ICTEDCT-CBIR model is compared with the CTDCIRS method [5], Jhanwar et al. [41], Hung and Dai’s [40] and Rao et al. [21] methods (as will be illustrated in Table 1 and Fig. 8, Fig. 9a and b, respectively). Moreover, experimental results have also been

Conclusions

In this paper, a novel Integrated Curvelet-based image retrieval scheme (ICTEDCT-CBIR) has been introduced, using Enhanced Dominant Color features and Texture analysis. The first contribution of the paper is the efficient integration of Curvelet Multiscale ridgelets of spatial band pass filtering with Region-based vector codebook Subband Clustering (RBSC) for dominant colors extraction and texture analysis. The second contribution is the design of an enhanced Region-based vector codebook

References (44)

Rui Min et al.
Effective image retrieval using dominant color descriptor and fuzzy support vector machine
Pattern Recog
(2009)
Jun Zhang et al.
Series feature aggregation for content-based image retrieval
J Comput Electr Eng
(2010)
N. Yang et al.
A fast MPEG-7 dominant color extraction with new similarity measure for image retrieval
J Vis Commun Image Represent
(2008)
P.W. Huang et al.
Image retrieval by texture similarity
Pattern Recogn
(2003)
N. Jhanwar et al.
Content based image retrieval using motif co-occurrence matrix
Image Vis Comput
(2004)
H.B. Kekre et al.
Walsh transform over row mean and column mean using image fragmentation and energy compaction for image retrieval
Int J Comput Sci Eng (IJCSE)
(2010)
Kekre HB, Thepade SD, Parkar A. A comparison of Kekre’s fast search and exhaustive search for various grid sizes used...
Dr.H.B. Kekre et al.
energy compaction and image splitting for image retrieval using kekre transform over row and column feature vectors
IJCSNS Int J Comput Sci Network Secur
(2010)
M. BabuRao et al.
CTDCIRS: content based image retrieval system based on dominant color and texture features
Int J Comput Appl (IJCSs)
(2011)
C. Kavitha et al.
Image retrieval based on color and texture features of the image sub-blocks
Int J Eng Sci Technol (IJEST)
(2011)

H. Tamura et al.

Texture features corresponding to visual perception

IEEE Trans Syst, Man Cybern

(1978)

B.S. Manjunath

Color and texture descriptors

IEEE Trans CSVT

(2001)

F. Liu et al.

Periodicity, directionality, and randomness: wold features for image modeling and retrieval

IEEE Trans Pattern Anal Mach Intell

(1996)

Huang Jing, Kumar S Ravi, Mitra Mandar, Zhu Wei-Jing, Zabi Ramin. Image indexing using color correlograms. In: Computer...

J. Engel

The multiresolution histogram

Metrika

(1997)

E. Hadidemetriou et al.

Multiresolution histograms and their use for recognition

IEEE Trans Pattern Anal Mach Intell

(2004)

J. Starck et al.

The curvelet transform for image denoising

IEEE Trans Image Process

(2002)

E. Candes

Fast discrete curvelet transforms

Multiscale Model Simul

(2006)

Chen L, Lu G, Zhang DS. Effects of different gabor filter parameters on image retrieval by texture. In: Proceedings of...

B.S. Manjunath et al.

Texture features for browsing and retrieval of large image data

IEEE Trans Pattern Anal Mach Intell

(1996)

Suematsu N, et al. Region-based image retrieval using wavelet transform. In: Proceedings of the 15th international...

Bhagavathy S, Chhabra K. A wavelet-based image retrieval system. Technical report—ECE278A, Santa Barbara: Vision...

Cited by (74)

A DCT probability histogram-based ROI features for content-based natural and medical image retrieval applications
2023, Digital Signal Processing: A Review Journal
The effectiveness of a Content-based Image Retrieval (CBIR) system depends on selecting representative and salient visual features. Capturing the salient image features for forming a feature vector is an essential part of the CBIR process, as it can significantly affect retrieval efficiency. Most images contain a visually prominent region that perceives maximum image semantic information, known as Regions-of-Interest (ROIs). This paper suggests a new CBIR scheme that employs a Graph-based Visual Saliency (GBVS) map to extract the ROI of the image. Further, block-level Discrete Cosine Transforms (DCT) and subsequent histograms are computed for the DC and significant AC coefficients of the extracted ROI image. In addition, a low-dimensional feature vector is formed from the modified histograms. Subsequently, six statistical features from the background scene have also been computed for the final feature vector construction. As a result, the final feature vector contains ROI and background information in their proper proportion to improve retrieval performance. Moreover, the performance of the proposed CBIR scheme has been evaluated using one medical, three natural, two objects, one texture, and one natural heterogeneous image database. All these databases collectively contain 53732 different images from 200 image classes. The outcomes of the experiments demonstrate substantial improvement in the retrieval process with some existing related schemes. At the same time, it also shows that the performance of the proposed scheme is robust concerning all kinds of image databases.
A hybrid CBIR system using novel local tetra angle patterns and color moment features
2022, Journal of King Saud University - Computer and Information Sciences
The exponential growth of communal media platforms including Facebook and Twitter, and the accessibility of low-cost digital capturing devices have generated an enormous number of multimedia content including images. Effective handling of such massive image collection has boosted the development of content-based image retrieval (CBIR) approaches. Researchers have suggested both machine learning and non-learning-based techniques for CBIR. However, machine learning-based methods outperform the non-learning-based methods in the CBIR domain. The CBIR demands the development of reliable descriptors to attain the most appropriate images from the depository and better address the semantic gap problem. To counter these problems, we suggest a novel second-order Local Tetra Angle Patterns (LTAP) to better capture the texture features from the image. LTAPs are computed from adjacent pixels of 0°, 45°, 90°, and 135° using the second-order directional derivatives. Further, we propose a hybrid feature vector by concatenating LTAPs and RGB color features and using the genetic algorithm (GA) to select the finest appropriate features that enhance the image retrieval performance of our system. We employed our hybrid descriptor with the GA to optimize the support vector machine (SVM) for the image classification task and used the Chi-square quadratic distance measure to determine the resemblance between the query image and the images in the repository. Experimental results on three standard datasets including the Corel 1 k, Oxford flower, and CIFAR-10 indicate the effectiveness of the presented system over the contemporary CBIR methods.
A CBIR system based on saliency driven local image features and multi orientation texture features
2022, Journal of Visual Communication and Image Representation
In Content-based Image Retrieval (CBIR), the user provides the query image in which only a selective portion of the image carries the foremost vital information known as the object region of the image. However, the human visual system also focuses on a particular salient region of an image to instinctively understand its semantic meaning. Therefore, the human visual attention technique can be well imposed in the CBIR scheme. Inspired by these facts, we initially utilized the signature saliency map-based approach to decompose the image into its respective main object region (ObR) and non-object region (NObR). ObR possesses most of the vital image information, so block-level normalized singular value decomposition (SVD) has been used to extract salient features of the ObR. In most natural images, NObR plays a significant role in understanding the actual semantic meaning of the image. Accordingly, multi-directional texture features have been extracted from NObR using Gabor filter on different wavelengths. Since the importance of ObR and NObR features are not equal, a new homogeneity-based similarity matching approach has been devised to enhance retrieval accuracy. Finally, we have demonstrated retrieval performances using both the combined and distinct ObR and NObR features on seven standard coral, texture, object, and heterogeneous datasets. The experimental outcomes show that the proposed CBIR system has a promising retrieval efficiency and outperforms various existing systems substantially.
A hierarchical CBIR framework using adaptive tetrolet transform and novel histograms from color and shape features
2018, Digital Signal Processing: A Review Journal
Citation Excerpt :
COREL-1000 is a standard natural image dataset containing 10 classes of images with huge diversity which makes it suitable for comparison analysis. Here, Table 9, 10 and 11 are showing the comparison between the proposed hierarchical CBIR systems based on CST combination with other standard CBIR systems present in the recent time span like Rao et al. [57], Elalami et al. [58], Youssef et al. [59], Subrahmanyam et al. [60], M. ElAlami et al. [61], Zeng et al. [62], Guo et al. [63] and Lande et al. [64]. The best retrieval results for each image categories are shown in bold letters in these tables.
Content-Based Image Retrieval (CBIR) systems retrieve the most analogous images from the image database with respect to a given query image based on the texture, shape, and/or color image features. These three image features can be used alone for the image retrieval or also can be used together for the retrieval purpose. In hierarchical CBIR system, three image features are extracted in proper order to discard the irrelevant images in each hierarchy level for reducing the image search space. In this paper, the authors have proposed a three-level hierarchical CBIR system/framework where, each level of the hierarchy uses either texture, shape or color image features to reduce the size of the image database by discarding the irrelevant images and at final level of the hierarchy, it will extract the most analogous images from the reduced image database. We have used adaptive tetrolet transform to extract the texture features from the regions of interest of the images. To extract the shape features of the image, a novel edge joint histogram has been proposed which uses the orientation of the edge pixels and their distance from the origin together to create a novel joint histogram. For color feature extraction, another color channel correlation histogram has been introduced. The order of the three different feature extraction processes on each level of the hierarchy is not rigid because it is difficult to predict the proper order for the highest retrieval. In the experiment, we have considered all possible order of the texture, shape and color features for image retrieval process. The retrieval experiments have been carried out in six different types of standard image databases and results show that the performance of proposed CBIR system has been increased significantly as compared to the other state-of-arts CBIR systems.
A survey on image data analysis through clustering techniques for real world applications
2018, Journal of Visual Communication and Image Representation
Citation Excerpt :
Color histogram is a most common color feature extraction approach. It generates different histograms for similar image having different viewpoints with different lightening conditions [53]. Whereas, color representation through histogram is invariant under rotation and translation [54].
A huge amount of image data is being collected in real world sectors. Image data analytics provides information about important facts and issues of a particular domain. But, it is challenging to handle voluminous, unstructured and unlabeled image collection. Clustering provides groups of homogeneous unlabeled data. Therefore, it is used quite often to access the interesting data easily and quickly. Image clustering is a process of partitioning image data into clusters on the basis of similarities. Whereas, features extracted from images are used for the computation of similarities among them. In this paper, significant feature extraction approaches and clustering methods applied on the image data from nine important applicative areas are reviewed. Medical, 3D imaging, oceanography, industrial automation, remote sensing, mobile phones, security and traffic control are considered applicative areas. Characteristics of images, suitable clustering approaches for each domain, challenges and future research directions for image clustering are discussed.
Assimilation of Pair-Wise Comparison Method to Decide Weights of Features Based on the Content of Image
2023, International Journal of Information Technology and Decision Making

View all citing articles on Scopus

Sherin M. Youssef received her PhD degree from University of Nottingham (UK, 2004) in Intelligent distributed swarm intelligence and optimization systems. She received her Master degree (MSc) from University of Alexandria (Egypt, 1995) in Pattern recognition and machine intelligence. She is currently an associate professor and a senior lecturer in the department of computer engineering, college of engineering, Arab Academy for science and Technology (AAST). Her research interests lie in the areas of Artificial intelligence, Intelligent agent based systems, digital signal processing, image processing, Video surveillance systems, Image retrieval systems, network security, Multimedia systems, and biomedical engineering.

^☆: Reviews processed and recommended for publication to Editor-in-Chief by Deputy Editor Dr. Ferat Sahin.

View full text

ICTEDCT-CBIR: Integrating curvelet transform with enhanced dominant colors extraction and texture analysis for efficient content-based image retrieval☆

Abstract

Graphical abstract

Highlights

Introduction

Section snippets

Literature survey

Digital curvelet transform

The proposed ICTEDCT-CBIR image retrieval model

Similarity matching

Experimental results

Conclusions

Pattern Recog

J Comput Electr Eng

J Vis Commun Image Represent

Pattern Recogn

Image Vis Comput

Walsh transform over row mean and column mean using image fragmentation and energy compaction for image retrieval

Int J Comput Sci Eng (IJCSE)

energy compaction and image splitting for image retrieval using kekre transform over row and column feature vectors

IJCSNS Int J Comput Sci Network Secur

CTDCIRS: content based image retrieval system based on dominant color and texture features

Int J Comput Appl (IJCSs)

Image retrieval based on color and texture features of the image sub-blocks

Int J Eng Sci Technol (IJEST)

Texture features corresponding to visual perception

IEEE Trans Syst, Man Cybern

Color and texture descriptors

IEEE Trans CSVT

Periodicity, directionality, and randomness: wold features for image modeling and retrieval

IEEE Trans Pattern Anal Mach Intell

The multiresolution histogram

Metrika

Multiresolution histograms and their use for recognition

IEEE Trans Pattern Anal Mach Intell

The curvelet transform for image denoising

IEEE Trans Image Process

Fast discrete curvelet transforms

Multiscale Model Simul

Texture features for browsing and retrieval of large image data

IEEE Trans Pattern Anal Mach Intell