Chia-Hung Wei

Trademark image retrieval using synthetic features for describing global shape and interior structure

A trademark image retrieval (TIR) system is proposed in this work to deal with the vast number of trademark images in the trademark registration system. The proposed approach commences with the extraction of edges using the Canny edge detector, performs a shape normalisation procedure, and then extracts the global and local features. The global features capture the gross essence of the shapes while the local features describe the interior details of the trademarks. A two-component feature matching strategy is used to measure the similarity between the query and database images. The performance of the proposed algorithm is compared against four other algorithms.

A Content-Based Approach to Medical Image Database Retrieval

Content-based image retrieval (CBIR) makes use of image features, such as color and texture, to index images with minimal human intervention. Content-based image retrieval can be used to locate medical images in large databases. This chapter introduces a content-based approach to medical image retrieval. Fundamentals of the key components of content-based image retrieval systems are introduced first to give an overview of this area. A case study, which describes the methodology of a CBIR system for retrieving digital mammogram database, is then presented. This chapter is intended to disseminate the knowledge of the CBIR approach to the applications of medical image management and to attract greater interest from various research communities to rapidly advance research in this field.

Mammogram retrieval on similar mass lesions

Enormous numbers of digital mammograms have been produced in hospitals and breast screening centers. To exploit those valuable resources in aiding diagnoses and research, content-based mammogram retrieval systems are required to effectively access the mammogram databases. This paper presents a content-based mammogram retrieval system, which allows medical professionals to seek mass lesions that are pathologically similar to a given example. In this retrieval system, shape and margin features of mass lesions are extracted to represent the characteristics of mammographic lesions. To compare the similarity between the query example and any lesion within the databases, this study proposes a similarity measure scheme which involves the hierarchical arrangement of mammographic features and a weighting distance measure. This makes similarity measure of the retrieval system consistent with the way radiologists observe mass lesions. This study used the DDSM dataset to evaluate the effectiveness of the extracted shape feature and margin feature, respectively. Experimental results demonstrate that, when Zernike moments are used, round-shape masses are the most discriminative among four types of shape; the circumscribed-margin masses can be effectively discriminated among the four types of margins. Moreover, the result also shows that, when retrieving round-shape and circumscribed margin masses, this retrieval system can achieve the highest precision among all mass lesion types.

Mammogram retrieval through machine learning within BI-RADS standards

A content-based mammogram retrieval system can support usual comparisons made on images by physicians, answering similarity queries over images stored in the database. The importance of searching for similar mammograms lies in the fact that physicians usually try to recall similar cases by seeking images that are pathologically similar to a given image. This paper presents a content-based mammogram retrieval system, which employs a query example to search for similar mammograms in the database. In this system the mammographic lesions are interpreted based on their medical characteristics specified in the Breast Imaging Reporting and Data System (BI-RADS) standards. A hierarchical similarity measurement scheme based on a distance weighting function is proposed to model users perception and maximizes the effectiveness of each feature in a mammographic descriptor. A machine learning approach based on support vector machines and users relevance feedback is also proposed to analyze the users information need in order to retrieve target images more accurately. Experimental results demonstrate that the proposed machine learning approach with Radial Basis Function (RBF) kernel function achieves the best performance among all tested ones. Furthermore, the results also show that the proposed learning approach can improve retrieval performance when applied to retrieve mammograms with similar mass and calcification lesions, respectively.

Effective Extraction of Gabor Features for Adaptive Mammogram Retrieval

Breast cancer is one of the most common diseases among women. Content-based mammogram retrieval has been proposed to aid various medical procedures. To develop a content-based mammogram retrieval system, textural feature extraction is one of the crucial requirements. This study proposes a Gabor filtering method for the extraction of textural features, which firstly performs Gabor filtering on the underlying image, applies the physical properties of a probability wave to probability transformation and then computes features to describe the textural pattern of the mammogram. This study also proposes an adaptive strategy for feature selection, filter selection and feature weighting, which utilizes a users relevance feedback to reduce the redundancy in the representation and incorporates the users information needs in image retrieval. Experimental results show that hypothesis tests can effectively find discriminated features and this retrieval system can improve its performance through just a few rounds of relevance feedback.

Protection of digital mammograms on PACSs using data hiding techniques

Picture archiving and communication systems (PACS) are typical information systems, which may be undermined by unauthorized users who have illegal access to the systems. This article proposes a role-based access control framework comprising two main components – a content-based steganographic module and a reversible watermarking module, to protect mammograms on PACSs. Within this framework, the content-based steganographic module is to hide patients’ textual information into mammograms without changing the important details of the pictorial contents and to verify the authenticity and integrity of the mammograms. The reversible watermarking module, capable of masking the contents of mammograms, is for preventing unauthorized users from viewing the contents of the mammograms. The scheme is compatible with mammogram transmission and storage on PACSs. Our experiments have demonstrated that the content-based steganographic method and reversible watermarking technique can effectively protect mammograms at PACS.

Calcification descriptor and relevance feedback learning algorithms for content-based mammogram retrieval

In recent years a large number of digital mammograms have been generated in hospitals and breast screening centers. To assist diagnosis through indexing those mammogram databases, we proposed a content-based image retrieval framework along with a novel feature extraction technique for describing the degree of calcification phenomenon revealed in the mammograms and six relevance feedback learning algorithms, which fall in the category of query point movement, for improving system performance. The results show that the proposed system can reach a precision rate of 0.716 after five rounds of relevance feedback have been performed.

A general framework for content-based medical image retrieval with its application to mammograms

In the field of medical imaging, content-based image retrieval (CBIR) techniques are employed to aid radiologists in the retrieval of images with similar contents. However, CBIR methods are usually developed based on specific features of images so that those methods are not readily inter-applicable among different kinds of medical images. This work proposes a general CBIR framework in attempt to alleviate this limitation. The framework is consisted of two parts: image analysis and image retrieval. In the image analysis part, normal and abnormal regions of interest (ROIs) in a number of images are selected to form a ROI dataset. These two groups of ROIs are used to analyze 11 textural features based on gray level co-occurrence matrices. The multivariate T test is then applied to identify the features with significant discriminating power for inclusion in a feature descriptor. In the image retrieval part, each feature of the descriptor is normalized by clipping the values of the largest 5% of the same feature component, and then projecting each normalized feature onto the unit sphere. The L2 norm is then employed to determine the similarity between the query image and each ROI in the dataset. This system works in the manner of query-by-example (QBE). Query images were selected from different classes of abnormal ROIs. A maximum precision of 51% and a maximum recall of 19% were obtained. The averages of precision and recall are 49% and 18% in this experiment.

Estimation of breast density: An adaptive moment preserving method for segmentation of fibroglandular tissue in breast magnetic resonance images

PURPOSE Breast density has been found to be a potential indicator for breast cancer risk. The estimation of breast density can be seen as a segmentation problem on fibroglandular tissues from a breast magnetic resonance image. The classic moment preserving is a thresholding method, which can be applied to determine an appropriate threshold value for fibroglandular tissue segmentation. METHODS This study proposed an adaptive moment preserving method, which combines the classic moment preserving and a thresholding adjustment method. The breast MR images are firstly performed to extract the fibroglandular tissue from the breast tissue. The next step is to obtain the areas of the fibroglandular tissue and the whole breast tissue. Finally, breast density can be estimated for the given breast. RESULTS The Friedman test shows that the qualities of segmentation are insignificant with p<0.000 and Friedman chi-squared=1116.12. The Friedman test shows that there would be significant differences in the sum of the ranks of at least one segmentation method. Average ranks indicate that the performance of the four methods is ranked as adaptive moment preserving, fuzzy c-means, moment preserving, and Kapurs method in order. Among the four methods, adaptive moment preserving also achieves the minimum values of MAE and RMSE with 9.2 and 12. CONCLUSION This study has verified that the proposed adaptive moment preserving can identify and segment the fibroglandular tissues from the 2D breast MR images and estimate the degrees of breast density.

Plant Identification Through Images: Using Feature Extraction of Key Points on Leaf Contours

Premise of the study: Because plant identification demands extensive knowledge and complex terminologies, even professional botanists require significant time in the field for mastery of the subject. As plant leaves are normally regarded as possessing useful characteristics for species identification, leaf recognition through images can be considered an important research issue for plant recognition. Methods: This study proposes a feature extraction method for leaf contours, which describes the lines between the centroid and each contour point on an image. A length histogram is created to represent the distribution of distances in the leaf contour. Thereafter, a classifier is applied from a statistical model to calculate the matching score of the template and query leaf. Results: The experimental results show that the top value achieves 92.7% and the first two values can achieve 97.3%. In the scale invariance test, those 45 correlation coefficients fall between the minimal value of 0.98611 and the maximal value of 0.99992. Like the scale invariance test, the rotation invariance test performed 45 comparison sets. The correlation coefficients range between 0.98071 and 0.99988. Discussion: This study shows that the extracted features from leaf images are invariant to scale and rotation because those features are close to positive correlation in terms of coefficient correlation. Moreover, the experimental results indicated that the proposed method outperforms two other methods, Zernike moments and curvature scale space.