IMAGAPP 2010 Abstracts


Area 1 - Imaging Theory

Full Papers
Paper Nr: 13
Title:

FUZZY SEGMENTATION OF MASSES IN DIGITAL BREAST TOMOSYNTHESIS IMAGES BASED ON DYNAMIC PROGRAMMING

Authors:

Louis Apffel, Giovanni Palma, Serge Muller and Isabelle Bloch

Abstract: In this paper we propose a new fuzzy segmentation method to segment lesions in Digital Breast Tomosynthesis (DBT) datasets. In the proposed approach we model a contour as a path in the image. The optimal contour is defined as the path associated with a minimal cost, which is derived from the image content. Using this formalism we present several ways to alter this cost in order to extract several relevant contours from a single image. The set of contours is then used in the fuzzy contour framework to perform mass detection. The method has been tested on synthetic data as well as images containing lesions and provides promising results.

Short Papers
Paper Nr: 4
Title:

IMPACT OF A LOSSY IMAGE COMPRESSION ON PARAMETER ESTIMATION WITH PERIODIC ACTIVE THERMAL IMAGING

Authors:

Agnès Delahaies, David Rousseau, Laetitia Perez, Laurent Autrique and François Chapeau-Blondeau

Abstract: Periodic thermal imaging is a method of active thermography based on a periodic thermal stimulation of an inspected sample material and the analysis of its thermal response when a steady regime is reached. The original data, a sequence of images sampling the thermal response on a large number of periods, are usually stored in a raw format. For accurate exploitation of these measurements, the whole sequence of images requires a significant amount of storage space. In this report, we address the question of the lossy compression of these sequences of images when they are applied to perform physical parameter estimation. The study investigates the impact of lossy image compression on the performance of the physical parameter estimation procedure, and shows the possibility of preserving robust estimation with high compression rate. Perspectives and applications are then discussed. Performing good enough estimate of physical parameters with compressed images would permit the use of portable thermal cameras with limited resources in terms of data storage. This would enable the use of periodic active thermal imaging to perform relatively low cost embedded characterization of thermal properties of materials.

Paper Nr: 10
Title:

SOLVING THE T-JOINT PROBLEM IN RECONSTRUCTING 2-D OBJECTS

Authors:

C. Stringfellow, R. Simpson, K. Enloe, R. Krasniqi, T. Ngo, R. Keown and J. Hood

Abstract: This paper describes a solution to the T-joint problem in matching 2D fragments of an object. Matching fragments of an object is useful for solving puzzles or reassembling archaeological fragments. Many factors, such as the number of pieces and the complex shapes of pieces make this a difficult problem. Various approaches to this problem exist. This paper presents an approach to solving the T-joint problem, which comes up in assembling fragments. The work described in this paper starts with a 2D object that should be easy to extend to 3D problems.

Paper Nr: 14
Title:

GPU-ACCELERATED IMAGE RETEXTURING IN GRADIENT DOMAIN

Authors:

Ping Li, Hanqiu Sun and Jianbing Shen

Abstract: This paper presents the novel GPU-accelated image retexturing approach for both high and low dynamic range images using our newly invented fast NLM filtering. Integrating the fast Maclaurin polynomial kernel filter and the latest GPU-CUDA acceleration, our approach is able to produce real-time high quality retexturing for objects of the interest, while preserving the original shading and similar texture distortion. We apply our revised NLM filtering to the initial depth map to ensure smoothed depth field for retexturing. Our approach using GPU-based fast NLM filtering is designed in parallel, and easy to develop on latest GPUs. Our testing results have shown the efficiency and satisfactory performance using our approach.

Paper Nr: 24
Title:

A NEW IMAGE RE-COLORING METHOD

Authors:

Hye-Yoon Woo and Hang-Bong Kang

Abstract: In this paper, we propose a new method to transform colors in the image according to the user’s emotion. First, we labeled the images into four basic emotions such as happy, sad, fear and anger from the users’ evaluation. After that, we compute color histograms using 36 quantized HSV scheme. After that, we intersect color histograms and finally construct color templates for basic emotions. To reflect the user’s preferences in color perception for specific emotion, we designed interactive genetic algorithm. First, we generate first stage individuals using color templates from the input image. From them, we extract chromosomes for HSV color space and then generate next stage individuals using crossover and mutation operations. After a few number of iterations, we stop the process and extract personalized color templates from the finalized image. Using personalized color templates, the user can generate an affection-based color transformed image according to his own preference. Survey results show that desirable results are obtained from our proposed personalized color templates.

Posters
Paper Nr: 25
Title:

A NEW SCHEME FOR DUAL WATERMARKING USING DWT-PCA TECHNIQUE

Authors:

Marzieh Amini, Hamidreza sadreazami and Khashayar Yaghmaie

Abstract: In this paper, we propose a new scheme of dual watermarking by applying discrete wavelet transform and principal component analysis. For the embedding, we identify two best sub bands of wavelet transform by considering the intensity variance of each sub band. Two watermarks are embedded within the selected sub bands with respect to their principal components. Using dual watermark improves robustness and imperceptibility of algorithm and also takes advantage of using efficient bandwidth by embedding two watermarks into host image. Interleaving technique is used to improve imperceptibility of the algorithm. Experimental results show no visible difference between host and watermarked images. The proposed watermarking method is also robust to various attacks such as JPEG compression, cropping, histogram modification and gamma correction. This robustness is more noticeable when cropping or gamma correction applies to the watermark image.

Area 2 - Imaging Applications

Full Papers
Paper Nr: 12
Title:

DIMENSION REDUCTION BASED ON CENTROIDS FOR MULTIMODAL ANATOMICAL LANDMARK-BASED 3D/2D REGISTRATION OF CORONARY ANGIOGRAMS

Authors:

Klaus Drechsler and Cristina Oyarzun Laura

Abstract: We present an anatomical landmark-based rigid 3D/2D registration algorithm to register computed tomography angiography (CTA) datasets with coronary angiograms (CA) gathered during a cardiac catheterization. It has to solve for six transformation parameters (three rotation and three translation parameters). An exhaustive search in a six dimensional search space is usually computationally very expensive and algorithms using optimization strategies can get lost in local minima. We propose a method based on centroids to reduce search space from six to four dimensions. Modern C-Arm devices store a lot of information about the acquisition geometry that are used to further reduce the search space. We use this method to develop an efficient smart exhaustive search to solve for the six transformation parameters in a competitive time. With our method registration errors of < 2 mm are feasible. Execution times of < 1 sec. can be reached on a QuadCore CPU.

Short Papers
Paper Nr: 17
Title:

UTERINE FIBROID SEGMENTATION ON MRI BASED ON CHAN-VESE LEVEL SET METHOD AND SHAPE PRIOR MODEL

Authors:

Alireza Fallahi, Mohammad Pooyan, Hassan Hashemi, Hassan Khotanlou, Mohammad Ali Oghabian and Kavous Firuznia

Abstract: Uterine fibroid the most common benign tumor of the female pelvic affected 20%- 50% of the women in the world. The efficacy of medical treatment is gauged by shrinkage of the size of these tumors after surgery. Complex fibroids anatomy, nonhomogeneity region and missing boundary in some cases are a challenging task in the segmentation. In this paper, we present a method to robustly segment these fibroids on MRI and measure the volume. Our method is based on combination of two step Chan-Vese level set method and geometric shape prior model. With calculating an initial region inside the fibroid using Chan-Vese level sets method, rough segmentation obtained followed by a prior shape model. We found the algorithm efficient and that it has some good results.

Area 3 - Imaging Technologies

Short Papers
Paper Nr: 16
Title:

CONSTRUCTING THE HYBRID DITHERING MATRIX WITH EQUAL CLUSTERED DOT DENSITY

Authors:

Xu Guoliang and Tan Qingping

Abstract: Hybrid halftone has great strengths over amplitude and frequency modulated halftone in offset printing. Modern CTP systems employ hybrid dithering algorithm, because dithering algorithm is a very efficient halftone algorithm. The shapes, distribution and density of clustered dots in halftone images depend on hybrid dithering matrix. This research proposes a new method to construct a hybrid dithering matrix with equal clustered dot density, based on a new geometry subdivision derived from Delaunay Triangulation. The matrix helps increase clustered dot density and maintain constraints of the offset printing procedure. This paper also discusses the uniformity of expanded central point set and quality of quadrilaterals of the geometry subdivision by comparing PSNR and quality factor value with that of the previous work.

Paper Nr: 21
Title:

OVERLAPPED BLOCK MOTION COMPENSATION FOR FRAME RATE UP CONVERSION

Authors:

Suk Kim, Taeuk Jeong and Chulhee Lee

Abstract: Recently, motion compensated interpolation methods are used to generate new frames in the frame rate up conversion. However, they often yield undesirable blocking artifacts due to inaccurate motion estimation. In order to mitigate these artifacts, we propose a new motion-compensated frame rate up conversion algorithm with overlapped block motion estimation. The new interpolated frame is obtained by overlapped block motion compensation of multiple motion estimation between the current and previous frames. Experimental results show the proposed method produces better perceptual quality and outperforms conventional motion-compensated methods in terms of PSNR.