Showing papers by "Dimitris Maroulis published in 2012"

TND: A Thyroid Nodule Detection System for Analysis of Ultrasound Images and Videos

Abstract: In this paper, we present a computer-aided-diagnosis (CAD) system prototype, named TND (Thyroid Nodule Detector), for the detection of nodular tissue in ultrasound (US) thyroid images and videos acquired during thyroid US examinations. The proposed system incorporates an original methodology that involves a novel algorithm for automatic definition of the boundaries of the thyroid gland, and a novel approach for the extraction of noise resilient image features effectively representing the textural and the echogenic properties of the thyroid tissue. Through extensive experimental evaluation on real thyroid US data, its accuracy in thyroid nodule detection has been estimated to exceed 95%. These results attest to the feasibility of the clinical application of TND, for the provision of a second more objective opinion to the radiologists by exploiting image evidences.

Fuzzy binary patterns for uncertainty-aware texture representation

Abstract: The Local Binary Pattern (LBP) representation of textures has been proved useful for a wide range of pattern recognition applications, including texture segmentation, face detection, and biomedical image analysis. The interest of the research community in the LBP texture representation gave rise to plenty of LBP and other binary pattern (BP)-based variations. However, noise sensitivity is still a major concern to their applicability on the analysis of real world images. To cope with this problem we propose a generic, uncertainty-aware methodology for the derivation of Fuzzy BP (FBP) texture models. The proposed methodology assumes that a local neighbourhood can be partially characterized by more than one binary patterns due to noise-originated uncertainty in the pixel values. The texture discrimination capability of four representative FBP-based approaches has been evaluated on the basis of comprehensive classification experiments on three reference datasets of natural textures under various types and levels of additive noise. The results reveal that the FBP-based approaches lead to consistent improvement in texture classification as compared with the original BP-based approaches for various degrees of uncertainty. This improved performance is also validated by illustrative unsupervised segmentation experiments on natural scenes.

Entropy-based spatially-varying adjustment of active contour parameters

Abstract: Parameter adjustment is a crucial, open issue in active contour methodology Most state-of-the-art active contours are empirically adjusted on a trial and error basis Such an empirical approach lacks scientific foundation, leads to suboptimal segmentation results and requires technical skills from the end-user This work introduces a method for automatic adjustment of active contour parameters, which is based on image entropy In addition, instead of being uniform, the parameter values calculated are spatially-varying, so as to reflect textural variations over the image Experimental evaluation of the proposed method is conducted on thyroid US images, liver MRI images, as well as on real-world photographs The results indicate that the proposed method is capable of identifying plausible object boundaries, obtaining a segmentation quality which is comparable to the one obtained with empirical parameter adjustment Moreover, the applicability of the proposed method is not confined on a single active contour variation

A contourlet transform based algorithm for real-time video encoding

Abstract: In recent years, real-time video communication over the internet has been widely utilized for applications like video conferencing. Streaming live video over heterogeneous IP networks, including wireless networks, requires video coding algorithms that can support various levels of quality in order to adapt to the network end-to-end bandwidth and transmitter/receiver resources. In this work, a scalable video coding and compression algorithm based on the Contourlet Transform is proposed. The algorithm allows for multiple levels of detail, without re-encoding the video frames, by just dropping the encoded information referring to higher resolution than needed. Compression is achieved by means of lossy and lossless methods, as well as variable bit rate encoding schemes. Furthermore, due to the transformation utilized, it does not suffer from blocking artifacts that occur with many widely adopted compression algorithms. Another highly advantageous characteristic of the algorithm is the suppression of noise induced by low-quality sensors usually encountered in web-cameras, due to the manipulation of the transform coefficients at the compression stage. The proposed algorithm is designed to introduce minimal coding delay, thus achieving real-time performance. Performance is enhanced by utilizing the vast computational capabilities of modern GPUs, providing satisfactory encoding and decoding times at relatively low cost. These characteristics make this method suitable for applications like video-conferencing that demand real-time performance, along with the highest visual quality possible for each user. Through the presented performance and quality evaluation of the algorithm, experimental results show that the proposed algorithm achieves better or comparable visual quality relative to other compression and encoding methods tested, while maintaining a satisfactory compression ratio. Especially at low bitrates, it provides more human-eye friendly images compared to algorithms utilizing block-based coding, like the MPEG family, as it introduces fuzziness and blurring instead of artificial block artifacts.

Accurate segmentation of 2D-PAGE images

Abstract: Segmentation is an essential and crucial process in 2D-PAGE image analysis. Although several software packages and techniques have been developed and are broadly utilized in biology laboratories, none of them can optimally segment 2D-PAGE images. This paper presents an effective approach for 2D-PAGE image segmentation. This approach extends our previous work by improving the segmentation accuracy. Experiments were conducted on a set of synthetic 16-bit 2D-PAGE images containing ∼2000 spots of various intensities. To the best of our knowledge, this is the first attempt to evaluate a spot segmentation method on 2D-PAGE images using a set of synthetic images that provide the challenges of real ones; inhomogeneous background, noise and artifacts, as well as faint, saturated and highly overlapping protein spots. Experimental results demonstrate that the proposed approach outperforms state of the art software packages and techniques since it: 1) generates more accurate boundaries and 2) separates overlapping spots correctly.

2D-GE image analysis focusing on elimination of spurious spots

Abstract: 2D Gel Electrophoresis image analysis is widely recognized as one of the most crucial processes following a proteomic experiment. Amongst its stages, detection and segmentation are the most challenging ones. The available software packages and techniques fail to detect and segment some of the real spots while they often detect a vast number of spurious spots. In this paper, an original approach to analyze gel electrophoresis images is presented. The proposed approach is an improvement on our earlier version because it tackles the aforementioned deficiencies of existing software programs and techniques in a more efficient manner. The experiments conducted on a set of 16-bit 2D gel electrophoresis images, demonstrate that it outperforms state-of-the-art methods. Indeed, it detects more real protein spot, as well as eliminates spurious spots.