Content-based Image Retrieval Approach Research Articles

Enhancing CBIR Performance Using Evolutionary Algorithm-Assisted Significant Feature Selection: A Filter Approach

Efficient searching becomes essential for large image archive, with more and more digital images available on Internet. Consequently, content-based image retrieval (CBIR) has drawn widespread research attentiveness in the last decade in the field of image processing, pattern recognition and computer vision. CBIR approach boils down to two core problems: feature extraction followed by feature matching. Feature selection is a process that selects pertinent features as a subset of original extracted features. This paper presents six filter approaches for significant features’ selection: decision tree, relief, genetic algorithm (GA) with correlation based feature selection (CFS) as fitness function, particle swarm optimisation (PSO) with CFS as fitness function, exhaustive search and forward selection with CFS as attribute subset evaluator. Work is carried out on publicly available Corel data set images. Experimental results prove that the features selected from PSO and GA with CFS enhances CBIR performance both in terms of higher retrieval accuracy and reduced computational time.

Content‐based image retrieval using fuzzy class membership and rules based on classifier confidence

Content representation for images with well-defined inter-class boundaries in the feature space remains to be a difficult task. Simple distance-based retrieval (SDR) approaches those operate on the feature space for content-based image retrieval (CBIR) are, therefore claimed to be inefficient by many researchers. Different CBIR approaches have been proposed to surmount the drawbacks of SDR scheme. This study proposes a novel image retrieval scheme. In this scheme, effort is taken to reduce the overall search time of the recently proposed approach called ‘class membership-based retrieval’ (CMR). The proposed method identifies the confidence in the classification and limits the search to single output class and therefore, reduces the overall search time by 21.76% as compared to CMR. Quantitative methods are proposed to select various parameters used in the algorithm which were computed empirically in the case of earlier approach CMR. The computed parameters are validated using experimental results. The consistent behaviours of the proposed method and earlier methods used in the experiment are demonstrated using different feature sets and distance metrics. While the method can be used as a general purpose image retrieval system, experiment is performed on four texture databases wit different complexities in terms of size, number of texture classes and orientation.

An efficient image retrieval scheme for colour enhancement of embedded and distributed surveillance images

From the past few years, the size of the data grows exponentially with respect to volume, velocity, and dimensionality due to wide spread use of embedded and distributed surveillance cameras for security reasons. In this paper, we have proposed an integrated approach for biometric-based image retrieval and processing which addresses the two issues. The first issue is related to the poor visibility of the images produced by the embedded and distributed surveillance cameras, and the second issue is concerned with the effective image retrieval based on the user query. This paper addresses the first issue by proposing an integrated image enhancement approach based on contrast enhancement and colour balancing methods. The contrast enhancement method is used to improve the contrast, while the colour balancing method helps to achieve a balanced colour. Importantly, in the colour balancing method, a new process for colour cast adjustment is introduced which relies on statistical calculation. It adjusts the colour cast and maintains the luminance of the image. The integrated image enhancement approach is applied to the enhancement of low quality images produced by surveillance cameras. The paper addresses the second issue relating to image retrieval by proposing a content-based image retrieval approach. The approach is based on the three features extraction methods namely colour, texture and shape. Colour histogram is used to extract the colour features of an image. Gabor filter is used to extract the texture features and the moment invariant is used to extract the shape features of an image. The use of these three algorithms ensures that the proposed image retrieval approach produces results which are highly relevant to the content of an image query, by taking into account the three distinct features of the image and the similarity metrics based on Euclidean measure. In order to retrieve the most relevant images, the proposed approach also employs a set of fuzzy heuristics to improve the quality of the results further. The results show the proposed approaches perform better than the well-known existing approaches.

Fusing Heterogeneous Features From Stacked Sparse Autoencoder for Histopathological Image Analysis.

In the analysis of histopathological images, both holistic (e.g., architecture features) and local appearance features demonstrate excellent performance, while their accuracy may vary dramatically when providing different inputs. This motivates us to investigate how to fuse results from these features to enhance the accuracy. Particularly, we employ content-based image retrieval approaches to discover morphologically relevant images for image-guided diagnosis, using holistic and local features, both of which are generated from the cell detection results by a stacked sparse autoencoder. Because of the dramatically different characteristics and representations of these heterogeneous features (i.e., holistic and local), their results may not agree with each other, causing difficulties for traditional fusion methods. In this paper, we employ a graph-based query-specific fusion approach where multiple retrieval results (i.e., rank lists) are integrated and reordered based on a fused graph. The proposed method is capable of combining the strengths of local or holistic features adaptively for different inputs. We evaluate our method on a challenging clinical problem, i.e., histopathological image-guided diagnosis of intraductal breast lesions, and it achieves 91.67% classification accuracy on 120 breast tissue images from 40 patients.

Assessment of performance and reproducibility of applying a content-based image retrieval scheme for classification of breast lesions.

To develop a new computer-aided diagnosis (CAD) scheme using a content-based image retrieval (CBIR) approach for classification between the malignant and benign breast lesions depicted on the digital mammograms and assess CAD performance and reproducibility. An image dataset including 820 regions of interest (ROIs) was used. Among them, 431 ROIs depict malignant lesions and 389 depict benign lesions. After applying an image preprocessing process to define the lesion center, two image features were computed from each ROI. The first feature is an average pixel value of a mapped region generated using a watershed algorithm. The second feature is an average pixel value difference between a ROI's center region and the rest of the image. A two-step CBIR approach uses these two features sequentially to search for ten most similar reference ROIs for each queried ROI. A similarity based classification score was then computed to predict the likelihood of the queried ROI depicting a malignant lesion. To assess the reproducibility of the CAD scheme, we selected another independent testing dataset of 100 ROIs. For each ROI in the testing dataset, we added four randomly queried lesion center pixels and examined the variation of the classification scores. The area under the ROC curve (AUC) = 0.962 ± 0.006 was obtained when applying a leave-one-out validation method to 820 ROIs. Using the independent testing dataset, the initial AUC value was 0.832 ± 0.040, and using the median classification score of each ROI with five queried seeds, AUC value increased to 0.878 ± 0.035. The authors demonstrated that (1) a simple and efficient CBIR scheme using two lesion density distribution related features achieved high performance in classifying breast lesions without actual lesion segmentation and (2) similar to the conventional CAD schemes using global optimization approaches, improving reproducibility is also one of the challenges in developing CAD schemes using a CBIR based regional optimization approach.

Color image representation using invariant exponent moments

In order to retrieve an image from a large image database, the descriptor should be invariant to geometric transformations (e.g., rotation, scaling, and translation). It must also have enough discriminating power and immunity to noise for retrieval from a large image database. The Exponent moments (EMs) descriptor has many desirable properties such as expression efficiency, robustness to noise, geometric invariance, fast computation, and multi-level representation. In this paper, we analyze the rotation, scaling, and translation (RST) invariant property of EMs, and propose a content-based image retrieval approach using invariant EMs. Experimental results show that the EMs can be used as an effective descriptor of global image content, and the proposed retrieval approach yields higher retrieval accuracy than some current state-of-the-art retrieval methods.

Image classification and retrieval using optimized Pulse-Coupled Neural Network

Content-Based Image Retrieval (CBIR) has become a powerful tool that is used in many image applications and search engines. Thus, many techniques and approaches for CBIR were developed in literature. The CBIR approach works on the visual features of the image rather than a descriptive text. Therefore, it provides more effective and efficient retrieval. On the other hand, PCNN has proved its efficiency as an image processing tool for various tasks such as image segmentation and recognition, feature extraction, edge and object detection. This article introduces a technique for content-based image classification and retrieval using PCNN. The proposed technique uses an optimized Pulse-Coupled Neural Network (PCNN) to extract the visual features of the image in a form of a numeric vector called image signature. An optimization mechanism was applied to the PCNN parameters in order to improve the signature quality. Thus improving the classification and retrieval results. Additionally, it employs the K-Nearest Neighbor (K-NN) algorithm for classification and matching. By applying classification before retrieval, the number of images in the search space is optimized to include one category instead of multiple categories. Moreover, we developed a CBIR prototype to validate our technique. The results show that our technique can retrieve and classify images efficiently. Furthermore, we evaluated our prototype against one of the widely used techniques and it was proven that the proposed technique can enhance the search results and improve the accuracy by 3.5%.

Content Based Image Retrieval Using Colour Strings Comparison

Content Based Image Retrieval (CBIR) is a technique that enables a user to extract an image based on a query, from a database containing a large amount of images. A very fundamental issue in designing a content based image retrieval system is to select the image features that best represent the image contents in a database. In this paper, our proposed method mainly concentrated on database classification and efficient image representation. We present a method for content based image retrieval based on support vector machine classifier. In this method the feature extraction was done based on the colour string coding and string comparison. We succeed in transferring the images retrieval problem to strings comparison. Thus the computational complexity is decreases obviously. The image database used in our experiment contains 1800 colour images from Corel photo galleries. This CBIR approach has significantly increased the accuracy in obtaining results for image retrieval.

Image Retrieval: Information and Rough Set Theories

Abstract In this research paper, we propose novel features based on information theory for image retrieval. We propose the novel concept of “probabilistic filtering”. We propose a hybrid approach for image retrieval that combines annotation approach with content based image retrieval approach. Also rough set theory is proposed as a tool for audio/video object retrieval from multi-media databases.

Content-based image retrieval approaches to interpret ground penetrating radar data

This paper presents a new data processing algorithm to interpret ground penetrating radar (GPR) data for quantification of railroad ballast fouling conditions. The algorithm is based on the observation that different fouling levels generate different textures in the GPR images. The algorithm was designed following the content-based image retrieval procedure, which includes two steps: feature extraction and similarity measurement. First, texture feature was extracted using discrete wavelet transform. Second, similarity measurement was performed. Laboratory GPR data were used to evaluate the accuracy of the algorithm. The accuracy was 93%, which demonstrated the effectiveness of the algorithm.

Content Based Image Retrieval Approach in Creating an Effective Feature Index for Lung Nodule Detection with the Inclusion of Expert Knowledge and Proven Pathology

The paper investigates four major issues in the active field of lung computer aided diagnosis (CAD) using content-based image retrieval (CBIR), which are: creating an efficient feature index for lung nodules for similarity measures, database creation of nodules with proven pathology, robust CBIR system and present a self-diagnosing environment to assist the physician in taking the right decision at right time. The results definitely improves the radiologists performance of detecting suspicious nodules based on the ground truth prepared. CBIR has been implemented to expand the small ground truth of 17 nodules to ground truth of 114 nodules based on available biopsy report. Nine out of 83 different extracted features have been considered as the best discriminating features to classify the lung nodules in three classes: Malignant, Benign and Metastasis. LIDC database has been analysed and achieved an average precision of 92.8% , mean average precision (MAP) of 82% at recall 0.1 and an average precision of 88% with PGIMER, Chandigarh. Results in this paper also indicate that the unnecessary biopsies can be avoided as the results are having few number of false positives which can directly increase the specificity of the proposed research. Keywords: CBIR, CAD, classification, lung cancer, nodules, LIDC.

Content-based image retrieval using high-dimensional information geometry

In this paper, a new content-based image retrieval approach is proposed based on high-dimensional information theory. The proposed approach overcomes the disadvantages of the current content-based image retrieval algorithms that suffer from the semantic gap. First, we present a new multidimensional information space’s vector angle cosine algorithm of high-dimensional geometry, then, we provide a detailed description of our images retrieval method including proposal of an overlapping image block method and definition of a similarity degree between images on the non-dimensional information subspaces. Finally, experimental results show the higher retrieval efficiency of the proposed algorithm.

Content Based Image Retrieval Using Spatial Features

The field of image retrieval technologies have enabled the creation of large image datasets. It is necessary to develop appropriate information systems to efficiently manage these collections. An image retrieval system is a computer system for searching, browsing and retrieving images from a large database of images.An approach known as Content-Based Image Retrieval (CBIR) systems used to retrieve images similar to a user-defined specification or pattern (image, sketch etc). Their goal is to support image retrieval based on content properties (colour, texture and shape), usually encoded into feature vectors. CBIR approach is an automatic retrieval process, instead of the traditional keyword-based approach, which usually is very time-consuming. The CBIR technology has been used in several applications such as, digital libraries, crime prevention, medicine, historical research, fingerprint identification, biodiversity information systems.

An Approach for the Cross Modality Content-Based Image Retrieval between Different Image Modalities

CBIR is an effective tool to search and extract image contents in a large remote sensing image database queried by an operator or end user. However, as imaging principles are different by sensors, their visual representation thus varies among image modality type. Considering images of various modalities archived in the database, image modality difference has to be tackled for the successful CBIR implementation. However, this topic has been seldom dealt with and thus still poses a practical challenge. This study suggests a cross modality CBIR (termed as the CM-CBIR) method that transforms given query feature vector by a supervised procedure in order to link between modalities. This procedure leverages the skill of analyst in training steps after which the transformed query vector is created for the use of searching in target images with different modalities. Current initial results show the potential of the proposed CM-CBIR method by delivering the image content of interest from different modality images. Despite its retrieval capability is outperformed by that of same modality CBIR (abbreviated as the SM-CBIR), the lack of retrieval performance can be compensated by employing the user's relevancy feedback, a conventional technique for retrieval enhancement.

Dual-Force ISOMAP: A New Relevance Feedback Method for Medical Image Retrieval

With great potential for assisting radiological image interpretation and decision making, content-based image retrieval in the medical domain has become a hot topic in recent years. Many methods to enhance the performance of content-based medical image retrieval have been proposed, among which the relevance feedback (RF) scheme is one of the most promising. Given user feedback information, RF algorithms interactively learn a user’s preferences to bridge the “semantic gap” between low-level computerized visual features and high-level human semantic perception and thus improve retrieval performance. However, most existing RF algorithms perform in the original high-dimensional feature space and ignore the manifold structure of the low-level visual features of images. In this paper, we propose a new method, termed dual-force ISOMAP (DFISOMAP), for content-based medical image retrieval. Under the assumption that medical images lie on a low-dimensional manifold embedded in a high-dimensional ambient space, DFISOMAP operates in the following three stages. First, the geometric structure of positive examples in the learned low-dimensional embedding is preserved according to the isometric feature mapping (ISOMAP) criterion. To precisely model the geometric structure, a reconstruction error constraint is also added. Second, the average distance between positive and negative examples is maximized to separate them; this margin maximization acts as a force that pushes negative examples far away from positive examples. Finally, the similarity propagation technique is utilized to provide negative examples with another force that will pull them back into the negative sample set. We evaluate the proposed method on a subset of the IRMA medical image dataset with a RF-based medical image retrieval framework. Experimental results show that DFISOMAP outperforms popular approaches for content-based medical image retrieval in terms of accuracy and stability.

Personal Clothing Retrieval on Photo Collections by Color and Attributes

Automatic personal clothing retrieval on photo collections, i.e., searching the same clothes worn by the same person, is not a trivial problem as photos are usually taken under completely uncontrolled realistic imaging conditions. Typically, the captured clothing images have large variations due to geometric deformation, occlusion, cluttered background, and photometric variability from illumination and viewpoint, which pose significant challenges to text-based or reranking-based visual search methods. In this paper, a novel framework is presented to tackle these issues by leveraging low-level features (e.g., color) and high-level features (attributes) of clothing. First, a content-based image retrieval (CBIR) approach based on the bag-of-visual-words (BOW) model is developed as our baseline system, in which a codebook is constructed from extracted dominant color patches. A reranking approach is then proposed to improve search quality by exploiting clothing attributes, including the type of clothing, sleeves, patterns, etc. Compared to low-level features, the attributes have better robustness to clothing variations, and carry semantic meanings as high-level image representations. Different visual attribute detectors are learned from large amounts of training data to extract the corresponding attributes. The construction of codebook and building of attribute classifiers are conducted offline, which leads to fast online search performance. Extensive experiments on photo collections show that the reranking algorithm based on attribute learning significantly improves retrieval performance in combination with the proposed baseline. Even our color-based baseline alone outperforms the previous CBIR-based search approaches. The experiments also demonstrate that our approach is robust to large variations of images taken in unconstrained environment.

Lung Segmentation in Chest Radiographs Using Anatomical Atlases With Nonrigid Registration

The National Library of Medicine (NLM) is developing a digital chest X-ray (CXR) screening system for deployment in resource constrained communities and developing countries worldwide with a focus on early detection of tuberculosis. A critical component in the computer-aided diagnosis of digital CXRs is the automatic detection of the lung regions. In this paper, we present a nonrigid registration-driven robust lung segmentation method using image retrieval-based patient specific adaptive lung models that detects lung boundaries, surpassing state-of-the-art performance. The method consists of three main stages: 1) a content-based image retrieval approach for identifying training images (with masks) most similar to the patient CXR using a partial Radon transform and Bhattacharyya shape similarity measure, 2) creating the initial patient-specific anatomical model of lung shape using SIFT-flow for deformable registration of training masks to the patient CXR, and 3) extracting refined lung boundaries using a graph cuts optimization approach with a customized energy function. Our average accuracy of 95.4% on the public JSRT database is the highest among published results. A similar degree of accuracy of 94.1% and 91.7% on two new CXR datasets from Montgomery County, MD, USA, and India, respectively, demonstrates the robustness of our lung segmentation approach.

Content-based Search for Image Retrieval

In this paper, a content-based image retrieval approach is presented for effective searching. The proposed approach uses two or more types of query for accessing images, textual annotation associated with images and visual appearance, such as colour, texture and positional features of objects in sample images. One can first place a keyword-based query, and then the desired images are retrieved by visual content-based query. The proposed retrieval approach shows clear improvements over competing approaches in terms of retrieval accuracy and visual inspection using Corel gallery and WWW images.

A New CBIR Approach for the Annotation of Medical Images

In the medical field, images, and especially digital images, are produced in ever increasing quantities and used for diagnostics and therapy. Imaging has occupied a huge role in the management of patients, whether hospitalized or not. This gave birth of the annotation of medical image process. The annotation is intended to image analysis and solve the problem of semantic gap. Physicians and radiologists feel better while using annotation techniques for faster making decision and giving solutions to patients in a faster and more accurate way. However, medical images annotation still a hard task specially the process based Content-based image retrieval (CBIR). Recently, advances in Content Based Image Retrieval prompted researchers towards new approaches in information retrieval for image databases. In medical applications it already met some degree of success in constrained problems. For this reason, we focus in this paper on presenting to provide an efficient semi-automatic tool which is used for efficient medical image retrieval from a huge content of medical image database and which is used for further medical diagnosis purposes for the new image annotation, because, efficient content-based image Retrieval in the medical domain is still a challenging problem. The goal of this work is to propose an approach able to compute similarity between a new medical image and old stored images. The annotator has to choose then one of the similar images and annotations related to the selected one are assigned to the new one. The idea is to apply an edge detector algorithm (Sobel algorithm) to the image and extract features from the filtered image by a color histogram. The edge to the image become likes Finger print to a human in our work. It is a search based edge. Edge representation of an image drastically reduces the amount of data to be processed, yet it retains important information about the shapes of objects in the scene. Edges in images constitute an important feature to represent their content and extraction features from filtered image improve searching of similar images, and keeping in the same time the properties of each image. The similarity measurement between images is developed based the Euclidean distance. The method can answer queries by example. The efficiency and performance of the presented method has been evaluated using the precision and the recall. The results of our experiments show high percentage of success, which is satisfactory.

Comparing the impact of Clustering with Content Based Image Retrieval Approaches for Plant Identification

Contour Based retrieval of images is an active and challenging field of research. Â Among various parameters available for contour based image retrieval, shape is considered an important aspect because it is closest to the human perception. Most of the shape based image retrieval methods require large processing time for generating accurate results due to huge database. To reduce the search time, we have divided the database into clusters on the basis of eccentricity of leaf using K-Means approach. After making the clusters, different contour based approaches are applied for leaf/plant identification and results are compared. Â The leaf image is processed to generate feature vectors which are stored in database. Â We have used Swedish leaf image database (SLID) consisting of 15 species with 75 leaves per class and total of 1125 leaf images. In this paper, we compare results of contour based retrieval approaches with and without clustering. From these results, it is found that by incorporating clustering, performance of contour based retrieval approaches remains same but retrieval time is reduced.