Medical image classification using an efficient data mining technique

Data mining refers to extracting or "mining" knowledge from large amounts of data. It is an increasingly popular field that uses statistical, visualization, machine learning, and other data manipulation and knowledge extraction techniques aimed at gaining an insight into the relationships and patterns hidden in the data. Availability of digital data within picture archiving and communication systems raises a possibility of health care and research enhancement associated with manipulation, processing and handling of data by computers.That is the basis for computer-assisted radiology development. Further development of computer-assisted radiology is associated with the use of new intelligent capabilities such as multimedia support and data mining in order to discover the relevant knowledge for diagnosis. It is very useful if results of data mining can be communicated to humans in an understandable way. In this paper, we present our work on data mining in medical image archiving systems. We investigate the use of a very efficient data mining technique, a decision tree, in order to learn the knowledge for computer-assisted image analysis. We apply our method to the classification of x-ray images for lung cancer diagnosis. The proposed technique is based on an inductive decision tree learning algorithm that has low complexity with high transparency and accuracy. The results show that the proposed algorithm is robust, accurate, fast, and it produces a comprehensible structure, summarizing the knowledge it induces.

Medical image retrieval with query-dependent feature fusion based on one-class SVM

Due to the huge growth of the World Wide Web, medical images are now available in large numbers in online repositories, and there exists the need to retrieval the images through automatically extracting visual information of the medical images, which is commonly known as content-based image retrieval (CBIR). Since each feature extracted from images just characterizes certain aspect of image content, multiple features are necessarily employed to improve the retrieval performance. Meanwhile, experiments demonstrate that a special feature is not equally important for different image queries. Most of existed feature fusion methods for image retrieval only utilize query independent feature fusion or rely on explicit user weighting. In this paper, we present a novel query dependent feature fusion method for medical image retrieval based on one class support vector machine. Having considered that a special feature is not equally important for different image queries, the proposed query dependent feature fusion method can learn different feature fusion models for different image queries only based on multiply image samples provided by the user, and the learned feature fusion models can reflect the different importances of a special feature for different image queries. The experimental results on the IRMA medical image collection demonstrate that the proposed method can improve the retrieval performance effectively and can outperform existed feature fusion methods for image retrieval.

QDFA : query-dependent feature aggregation for medical image retrieval

We propose a novel query-dependent feature aggregation (QDFA) method for medical image retrieval. The QDFA method can learn an optimal feature aggregation function for a multi-example query, which takes into account multiple features and multiple examples with different importance. The experiments demonstrate that the QDFA method outperforms three other feature aggregation methods.

A new query dependent feature fusion approach for medical image retrieval based on one-class SVM

With the development of the internet, medical images are now available in large numbers in online repositories, and there exists the need to retrieval the medical images in the content-based ways through automatically extracting visual information of the medical images. Since a single feature extracted from images just characterizes certain aspect of image content, multiple features are necessarily employed to improve the retrieval performance. Furthermore, a special feature is not equally important for different image queries since a special feature has different importance in reflecting the content of different images. However, most existed feature fusion methods for image retrieval only utilize query independent feature fusion or rely on explicit user weighting. In this paper, based on multiply query samples provided by the user, we present a novel query dependent feature fusion method for medical image retrieval based on one class support vector machine. The proposed query dependent feature fusion method for medical image retrieval can learn different feature fusion models for different image queries, and the learned feature fusion models can reflect the different importance of a special feature for different image queries. The experimental results on the IRMA medical image collection demonstrate that the proposed method can improve the retrieval performance effectively and can outperform existed feature fusion methods for image retrieval.

A new re-ranking method using enhanced pseudo-relevance feedback for content-based medical image retrieval

We propose a novel re-ranking method for content-based medical image retrieval based on the idea of pseudo-relevance feedback (PRF). Since the highest ranked images in original retrieval results are not always relevant, a naive PRF based re-ranking approach is not capable of producing a satisfactory result. We employ a two-step approach to address this issue. In step 1, a Pearson's correlation coefficient based similarity update method is used to re-rank the high ranked images. In step 2, after estimating a relevance probability for each of the highest ranked images, a fuzzy SVM ensemble based approach is adopted to re-rank the images. The experiments demonstrate that the proposed method outperforms two other re-ranking methods.

Medical image retrieval based on unclean image bags

Traditional content-based image retrieval (CBIR) scheme with assumption of independent individual images in large-scale collections suffers from poor retrieval performance. In medical applications, images usually exist in the form of image bags and each bag includes multiple relevant images of the same perceptual meaning. In this paper, based on these natural image bags, we explore a new scheme to improve the performance of medical image retrieval. It is feasible and efficient to search the bag-based medical image collection by providing a query bag. However, there is a critical problem of noisy images which may present in image bags and severely affect the retrieval performance. A new three-stage solution is proposed to perform the retrieval and handle the noisy images. In stage 1, in order to alleviate the influence of noisy images, we associate each image in the image bags with a relevance degree. In stage 2, a novel similarity aggregation method is proposed to incorporate image relevance and feature importance into the similarity computation process. In stage 3, we obtain the final image relevance in an adaptive way which can consider both image bag similarity and individual image similarity. The experiments demonstrate that the proposed approach can improve the image retrieval performance significantly.

A noisy-smoothing relevance feedback method for content-based medical image retrieval

In this paper, we address a new problem of noisy images which present in the procedure of relevance feedback for medical image retrieval. We concentrate on the noisy images, caused by the users mislabeling some irrelevant images as relevant ones, and a noisy-smoothing relevance feedback (NS-RF) method is proposed. In NS-RF, a two-step strategy is proposed to handle the noisy images. In step 1, a noisy elimination algorithm is adopted to identify and eliminate the noisy images. In step 2, to further alleviate the influence of noisy images, a fuzzy membership function is employed to estimate the relevance probabilities of retained relevant images. After noisy handling, the fuzzy support vector machine, which can take into account different relevant images with different relevance probabilities, is adopted to re-rank the images. The experimental results on the IRMA medical image collection demonstrate that the proposed method can deal with the noisy images effectively.

An image retrieval system based on fractal dimension

This paper presents a new kind of image retrieval system which obtains the feature vectors of images by estimating their fractal dimension; and at the same time establishes a tree-structure image database. After preprocessing and feature extracting, a given image is matched with the standard images in the image database using a hierarchical method of image indexing.

Image retrieval based on the combination of color and keyword

In this paper, color information and keyword information are combined for image retrieval. In detail, each image is divided into several blocks and then the color histograms of each block are derived. Users could feed back some annotations represented by keywords. Then, the keywords may spread in the image database so that both color-based and keyword-based retrieval could be utilized together. A prototype system shows that the proposed method is effective and efficient in performing image retrieval tasks.

Hybrid classification of pulmonary nodules

Automated classification of lung nodules is challenging because of the variation in shape and size of lung nodules, as well as their associated differences in their images. Ensemble based learners have demonstrated the potentialof good performance. Random forests are employed for pulmonary nodule classification where each tree in the forest produces a classification decision, and an integrated output is calculated. A classification aided by clustering approach is proposed to improve the lung nodule classification performance. Three experiments are performed using the LIDC lung image database of 32 cases. The classification performance and execution times are presented and discussed.

A neural network based human identification framework using ear images

This paper presents a framework that uses ear images for human identification. The framework makes use of Principal Component Analysis (PCA) for ear image feature extraction and Multilayer Feed Forward Neural Network for classification. Framework are proposed to improve recognition accuracy of human identification. The framework was tested on an ear image database to evaluate its reliability and recognition accuracy. The experimental results showed that our framework achieved higher stable recognition accuracy and over-performed other existing methods. The recognition accuracy stability and computation time with respect to different image sizes and factors were investigated thoroughly as well in the experiments.

Semantic photo synthesis

Composite images are synthesized from existing photographs by artists who make concept art, e.g., storyboards for movies or architectural planning. Current techniques allow an artist to fabricate such an image by digitally splicing parts of stock photographs. While these images serve mainly to “quickly”convey how a scene should look, their production is laborious. We propose a technique that allows a person to design a new photograph with substantially less effort. This paper presents a method that generates a composite image when a user types in nouns, such as “boat”and “sand.”The artist can optionally design an intended image by specifying other constraints. Our algorithm formulates the constraints as queries to search an automatically annotated image database. The desired photograph, not a collage, is then synthesized using graph-cut optimization, optionally allowing for further user interaction to edit or choose among alternative generated photos. An implementation of our approach, shown in the associated video, demonstrates our contributions of (1) a method for creating specific images with minimal human effort, and (2) a combined algorithm for automatically building an image library with semantic annotations from any photo collection.

BioKin: an ambulatory platform for gait kinematic and feature assessment

A platform to move gait analysis, which is normally restricted to a clinical environment in a well-equipped gait laboratory, into an ambulatory system, potentially in non-clinical settings is introduced. This novel system can provide functional measurements to guide therapeutic interventions for people requiring rehabilitation with limited access to such gait laboratories. BioKin system consists of three layers: a low-cost wearable wireless motion capture sensor, data collection and storage engine, and the motion analysis and visualisation platform. Moreover, a novel limb orientation estimation algorithm is implemented in the motion analysis platform. The performance of the orientation estimation algorithm is validated against the orientation results from a commercial optical motion analysis system and an instrumented treadmill. The study results demonstrate a root-mean-square error less than 4° and a correlation coefficient more than 0.95 when compared with the industry standard system. These results indicate that the proposed motion analysis platform is a potential addition to existing gait laboratories in order to facilitate gait analysis in remote locations.

Evaluation and comparison of 3D intervertebral disc localization and segmentation methods for 3D T2 MR data: A grand challenge.

The evaluation of changes in Intervertebral Discs (IVDs) with 3D Magnetic Resonance (MR) Imaging (MRI) can be of interest for many clinical applications. This paper presents the evaluation of both IVD localization and IVD segmentation methods submitted to the Automatic 3D MRI IVD Localization and Segmentation challenge, held at the 2015 International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI2015) with an on-site competition. With the construction of a manually annotated reference data set composed of 25 3D T2-weighted MR images acquired from two different studies and the establishment of a standard validation framework, quantitative evaluation was performed to compare the results of methods submitted to the challenge. Experimental results show that overall the best localization method achieves a mean localization distance of 0.8 mm and the best segmentation method achieves a mean Dice of 91.8%, a mean average absolute distance of 1.1 mm and a mean Hausdorff distance of 4.3 mm, respectively. The strengths and drawbacks of each method are discussed, which provides insights into the performance of different IVD localization and segmentation methods.