Search

This paper explores the potential use of WebRTC set of protocols for DICOM file exchange. We have developed a simple proof-of-concept peer-to-peer DICOM file-sharing web application based on a set of WebRTC protocols. Application performance is compared with contemporary DICOM applications and transfer protocols which showed that WebRTC has its place in the DICOM file-sharing domain.

With the growing number of face recognition applications in everyday life, image- and video-based recognition methods are becoming important research topic (Zhao et al., 2003). Effects of pose, illumination and expression are issues currently most studied in face recognition. So far, very little has been done to investigate the effects of compression on face recognition, even though the images are mainly stored and/or transported in a compressed format. Still-to-still image experimental setups are often researched, but only in uncompressed image formats. Still-to-video research (Zhou et al., 2003) mostly deals with issues of tracking and recognizing faces in a sense that still uncompressed images are used as a gallery and compressed video segments as probes. In this chapter we analyze the effects that standard image compression methods - JPEG (Wallace, 1991) and JPEG2000 (Skodras et al., 2001) - have on two well known subspace appearance-based face recognition algorithms: Principal Component Analysis - PCA (Turk & Pentland, 1991), Linear Discriminant Analysis - LDA (Belhumeur et al., 1996) and Independent Component Analysis - ICA (Bartlett et al., 2002). We use McNemar's hypothesis test (Beveridge et al., 2001 ; Delac et al., 2006) when comparing recognition accuracy in order to determine if the observed outcomes of the experiments are statistically important or a matter of chance. Following the idea of a reproducible research, a comprehensive description of our experimental setup is given, along with details on the choice of images used in the training and testing stage, exact preprocessing steps and recognition algorithms parameters setup. Image database chosen for the experiments is the grayscale portion of the FERET database (Phillips et al., 2000) and its accompanying protocol for face identification, including standard image gallery and probe sets. Image compression is performed using standard JPEG and JPEG2000 coder implementations and all experiments are done in pixel domain (i.e. the images are compressed to a certain number of bits per pixel and then uncompressed prior to use in recognition experiments). The recognition system's overall setup we test is twofold. In the first part, only probe images are compressed and training and gallery images are uncompressed (Delac et al., 2005). This setup mimics the expected first step in implementing compression in real-life face recognition applications: an image captured by a surveillance camera is probed to an existing high-quality gallery image. In the second part, a leap towards justifying fully compressed domain face recognition is taken by using compressed images in both training and testing stage (Delac, 2006). We will show that, contrary to common opinion, compression does not deteriorate performance but it even improves it slightly in some cases. We will also suggest some prospective lines of further research based on our findings.

For a number of years, various games have been used as an educational tool at different academic levels, especially in primary education. However, only recently games that teach coding and algorithmic thinking or even broader, computational thinking, emerged. Initiatives like Hour of Code and similar online activities or block-based programming games popularized the field, while at the same time, projects like CSUnplugged showed that the “idea of programming” can be learned even without using the computer. In this paper, we present our experience so far in creating and implementing different unplugged activities that teach students of different age levels, from early primary school to the university students, the simple programming concepts and algorithms. As a part of Science, Technology, Engineering, Mathematics and Computer Science (STEM- C) outreach program named SUZA - From school to science and the academic community at University of Zagreb Faculty of Electrical Engineering and Computing, Croatia, we conducted a number of game-like activities based on graph paper programming, block-based programming and using the existing board games. Although this is a work in progress project, the participant reactions so far, collected through interviews and questionnaires, indicated that the conducted activities were well accepted by students and their teachers. We also received a number of useful feedback comments and proposals, such as to expand the activities to include the homework part, or to make them more physical and relocate them outdoors.

This paper presents a review of recent advances in the development of methods for segmentation of the breast boundary and the pectoral muscle in mammograms. Regardless of improvement of imaging technology, accurate segmentation of the breast boundary and detection of the pectoral muscle are still challenging tasks for image processing algorithms. In this paper, we discuss problems related to mammographic image preprocessing and accurate segmentation. We review specific methods that were commonly used in most of the techniques proposed for segmentation of mammograms and discuss their advantages and disadvantages. Comparative analysis of the methods reported on is made difficult by variations in the datasets and procedures of evaluation used by the authors. We attempt to overcome some of these limitations by trying to compare methods which used the same dataset and have some similarities in approaches to the breast boundary segmentation and detection of the pectoral muscle. In this paper, we will address the most often used methods for segmentation such as thresholding, morphology, region growing, active contours, and wavelet filtering. These methods, or their combinations, are the ones most used in the last decade by the majority of work published in this image processing domain.

Digital radiology imaging systems, based on DICOM standard, are generating more and more images annually. Access to the images is possible through DIMSE network language or DICOM web methods. Peer-to-peer networks, on the other hand, have gained some interest in the DICOM community, but most have been surpassed by cloud computing methods. WebRTC seems like a potential technology for helping DICOM standard to gain P2P connectivity, which could help both business and scientific community easier access to the images repositories. In this paper, we examine all present technologies.

Today it is already widely accepted that out of school education exceeds formal education in content and knowledge and that it is not a plus, but a necessity. However, unlike formal education, educators working in informal settings are often volunteers or/and they do those activities on top of their daily jobs. The research questions we pose here are (i) what motivates people to volunteer in K-12 Science, Technology, Engineering, Mathematics and Computer Science (STEM-C) informal education, (ii) what could help to motivate people who are currently demotivated, and finally (iii) what we can do not only to attract new volunteers, but also to retain the current ones. We collected and analyzed opinions of faculty staff, students and volunteers involved in K-12 STEM-C outreach activities conducted at University of Zagreb Faculty of Electrical Engineering and Computing, Croatia where five years ago, we started our outreach program named SUZA — From school to science and the academic community. The results of our research study show a wide span of reasons why (and why not) people volunteer in our activities, together with their general attitudes toward K-12 STEM-C outreach activities. Although the results are mostly in line with research in the field, there are some specifics which could relate to specifics of volunteering in K-12 STEM-C fields and could benefit wider community.

In this paper, we present a new full-reference objective image quality measure—IQM2, based on structural similarity index and steerable pyramid wavelet transform. IQM2 is tested using different number of orientation kernels and seven subjective databases. Finally, IQM2 measure is compared with twelve commonly used full-reference objective measures. Results show that proposed IQM2 measure, using kernel with 2 orientations, provides good correlation with the results of subjective evaluation while keeping computational time lower than other similar performing objective measures.

Breast skin–air interface and pectoral muscle segmentation are usually first steps in all CAD applications on scanned as well as digital mammograms. Breast skin–air interface segmentation is much more difficult task when performed on scanned mammograms than on digital mammograms. In case of pectoral muscle segmentation, segmentation difficulty of analog and digital mammograms is usually similar. In this paper we present adaptive contrast enhancement method for breast skin–air interface detection which combines usage of adaptive histogram equalization method on small region of interest which contains actual edge and edge detection operators. Pectoral muscle detection method uses combination of contrast enhancement using adaptive histogram equalization and polynomial curvature estimation on selected region of interest. This method makes segmentation of very low contrast pectoral muscle areas possible because of estimation used to segment areas which have lower contrast difference than detection threshold.

Even students who are provided with currently best available education in Science, Technology, Engineering, Mathematics and Computer Science (STEM-C) fields are having problems coping with living in the modern world where technological advancements happen on a daily basis, let alone students coming from economically disadvantaged backgrounds. In a world where institutions offering formal education have to collaborate with individuals and groups of people interested in providing informal education, it is of a vital importance to set up good examples and share them with communities that are less experienced. In this paper we present how Student Research Development Center's (SRDC) best practices of establishing a pre-college pipeline for young economically disadvantage minority students, who are interested in STEM-C fields, from Puerto Rico are being transferred to Universidad Catolica de Nicaragua (UNICA). The goal of the paper is to show how we developed a partnership between Puerto Rico and Nicaragua, and used lessons learned in Puerto Rico to involve undergraduate and pre-college students from Nicaragua in research program using project based learning in STEM-C fields.

[1. Background and Motivation] This special issue of the Multimedia Tools and Applications Journal contains extended papers selected from the image and video processing sessions and multimedia sessions of ELMAR symposia held in Zadar, Croatia, in 2008 and 2009 (more information about ELMAR symposia can be found here: www.elmar-zadar.org). Eight selected papers provide a coverage of recent advances in image and video processing. Papers were selected by the Guest Editors of this special issue in September 2009, and the respective authors were invited to submit the extended versions of their papers, along with further research they conducted on the subject. // Multimedia is all around us and we are often reminded that we live in the age of multimedia. It is in products and services like digital TV, IPTV, video and image databases, social networking etc. Time between research and final consumer product has never been shorter and this is one of the reasons why multimedia is still an attractive and widely researched area. Given the wide area of multimedia tools and applications usage in real life, it is necessary to research and develop algorithms and methods in various fields, and this special issue reflects this fact in its wide range of topics covered. // [2. Scanning the Special Issue] The two papers that open this special issue deal with an important subject of video quality assessment. The first paper is "Comparison of HDTV Formats using Objective Video Quality Measures" by Emil Dumic, Sonja Grgic and Mislav Grgic. By combining subjective and objective quality measures the authors show that progressive scanning should be preferred to interlaced for all future HDTV emissions. The second paper is "Foveated Mean Squared Error - A Novel Video Quality Metric" where Snjezana Rimac-Drlje, Mario Vranjes and Drago Zagar propose a new video quality measure called Foveated Mean Squared Error (FMSE) that takes into account the non-uniform sampling corresponding to the visual information acquisition process at the human retina. FMSE shows high correlation to the subjective quality assessment. // Segmentation is the central theme of the next three papers in this special issue. In "Segmentation of Optic Disc in Retinal Images using an Improved Gradient Vector Flow Algorithm" by Huiyu Zhou, Gerald Schaefer, Tangwei Liu and Faquan Lin, a novel medical image segmentation algorithm is presented. It is based on gradient vector flow and is used for segmenting the optic disc in retinal images to support automatic screening. Michal Kawulok addressed the problem of precise skin segmentation for sign language purposes in his article entitled "Energy-based Blob Analysis for Improving Precision of Skin Segmentation". The presented approach uses blob analysis and takes into account skin pixel topology which is usually disregarded by other color-based skin detectors. The paper "Automatic MPEG-4 Sprite Coding - Comparison of Integrated Object Segmentation Algorithms", by Alexander Glantz, Andreas Krutz, Thomas Sikora, Paulo Nunes and Fernando Pereira, proposes an automatic object segmentation algorithm for video sequences with a moving camera towards automatic object representation for sprite coding as standardized in MPEG-4 Visual. In their experiments, the authors have showed that their algorithm performs comparable to manually segmented groundtruth masks, implying almost perfect segmentation. // The last three papers cover areas of content retrieval, watermarking and face recognition. In the paper entitled: "Interactive Search and Browsing Interface for Large-scale Visual Repositories", Kan Ren, Risto Sarvas and Janko Calic present a way in which a user can intuitively navigate through a large image and video collections. In the paper "Two Dimensional Time-Frequency Analysis based Eigenvalue Decomposition Applied to Image Watermarking", Srdjan Stankovic, Irena Orovic, Nikola Zaric and Cornel Ioana propose applying spatial-frequency analysis and eigenvalue decomposition to provide an efficient image watermarking method. The efficiency of the proposed method and its robustness in the presence of various attacks is proven in extensive reported experiments. The last paper of this special issue is "Aligned Texture Map Creation for Pose Invariant Face Recognition" by Antonio Rama, Francesc Tarrés and Jürgen Rurainsky. The paper deals with a novel face recognition method called Partial Principal Component Analysis. The method has been proved to be efficient in pose-varying recognition scenarios. The paper also presents a novel approach for the automatic creation of 180 degrees aligned cylindrical projected face images using different views, which are then used for training purposes in the presented face recognition method. // [Acknowledgements] The Guest Editors would like to thank all those who have helped to make this special issue possible, especially to ELMAR symposium organizers and committee members, all authors and the reviewers. The Guest Editors would like to gratefully acknowledge Prof. Borko Furht, Editor-in-Chief of the Multimedia Tools and Applications Journal, for giving them the opportunity to publish this special issue.

In this paper we describe a database of static images of human faces. Images were taken in uncontrolled indoor environment using five video surveillance cameras of various qualities. Database contains 4,160 static images (in visible and infrared spectrum) of 130 subjects. Images from different quality cameras should mimic real-world conditions and enable robust face recognition algorithms testing, emphasizing different law enforcement and surveillance use case scenarios. In addition to database description, this paper also elaborates on possible uses of the database and proposes a testing protocol. A baseline Principal Component Analysis (PCA) face recognition algorithm was tested following the proposed protocol. Other researchers can use these test results as a control algorithm performance score when testing their own algorithms on this dataset. Database is available to research community through the procedure described at www.scface.org .

Facial image processing and analysis methods have numerous applications in visual surveillance, computer and physical access control, and digital entertainment. Since facial images are commonly used on various identification documents, these methods have a great potential to become the next generation biometric technology. As facial images can be acquired even without subjects' cooperation, they constitute often the only biometric information available in some visual data databases. // Though there has been a great deal of progress in face detection and recognition in the last few years, many problems remain unsolved. Rigorous testing and verification on real-world datasets often reveals that today's facial image processing and analysis systems still lack the robustness required for reliable deployment. As an example, face detection must cope with many challenging problems, especially when dealing with outdoor illumination, pose variation with large rotation angles, low image quality, low resolution, occlusion, and background changes characteristic of complex real-life scenes. The design of face recognition algorithms that are effective over a wide range of viewpoints, complex outdoor lighting, occlusions, facial expressions and aging is still a major area of research. There is a continuous need to improve detection and recognition, mode and evaluate performance, and develop real-time biometric systems. // This special issue focuses on recent advances in facial image processing and analysis that address the above problems and explores emerging biometric applications. The purpose is to fill an existing gap in the scientific literature by presenting recent progress and to provide benchmarks and reference points for future developments. // The idea of putting together the Special Issue on Facial Image Processing and Analysis was suggested to the Editors-In-Chief by the Guest Editors in November 2006. In December 2006, the outline and schedule of the special issue were established by the Guest Editors, and the first call for papers was distributed through the Internet. // Since May to December 2007, forty-six manuscripts were submitted for possible inclusion in the special issue. Each one of the submitted manuscripts was reviewed by at least three experts in the field. Between December 2007 and March 2008, the first round of reviews was completed with thirty papers removed from further consideration. Between September and November 2008, revised versions of the remaining papers underwent an additional review by the Guest Editors who eventually selected twelve articles for the special issue. The final issue was finalized in February 2009. // The papers included in the special issue cover great diversity in the methods and applications proposed for facial image processing and analysis. It is our hope that the accepted papers deal with relevant practical issues and challenging problems that will prove valuable to both researchers and practitioners. // The special issue opens with a paper on "Precise Eye and Mouth Localization" by P. Campadelli, R. Lanzarotti, and G. Lipori. It presents a method which is based on support vector machines trained on optimally chosen Haar wavelet coefficients. In another paper on facial feature localization, entitled "Towards Practical Facial Feature Detection", M. Eckhardt, I. Fasel, and J. Movellan develop the idea of context dependent inference. Their real-time system first uses robust detectors to detect contexts in which target features occur and then employs refined detectors trained to localize the features given the detected context. // As the choice of a method for representing common patterns is a critical factor in the design of robust facial image detection and classification systems, M. Ashraful Amin and H. Yan present "An Empirical Study on the Characteristics of Gabor Representations for Face Recognition, " which compares the classification capability of different Gabor representations for human face recognition and introduces a cost-effective variant of Gabor feature extraction methods. In the next paper, entitled "Face Detection and Recognition Using Maximum Likelihood Classifiers on Gabor Graphs", M. Gunther and R. P. Wurtz propose an integrated face detection and recognition system that combines various classifiers and Gabor graphs. // Feature extraction is the central theme in "Subspaces versus Submanifolds: A Comparative Study in Small Sample Size Problem" by H. Huang, J. Li, and H. Feng. The authors compare a number of unsupervised, supervised, and kernel-based methods for face recognition. In another paper on feature extraction, "Robust Adapted Principal Component Analysis for Face Recognition", S. Chen, B. C. Lovell and T. Shan deal simultaneously with large variations in illumination, expression and pose using only a single gallery image per person. // In "Facial Biometrics Using Non-tensor Product Wavelet and 2D Discriminant Techniques", D. Zhang, X. You, P. Wang, S. N. Yanushkevich, and Y. Y. Tang propose a non-tensor product bivariate wavelet coupled with a modified two-dimensional linear discriminant technique that allows the detection of particular facial features in the high-frequency components. To compensate for pose changes and limited occlusion and distortions, F. Li and H. Wechsler propose "Face Authentication Using Recognition-by-Parts, Boosting and Transduction". This method involves feature selection of local patch instances including dimensionality reduction, exemplar-based clustering of patches into parts, and data fusion for matching using boosting driven by parts that play the role of weak-learners. // The next two papers focus on face analysis for digital video applications. In the paper entitled "Multi-Scale Dynamic Features Based Driver Fatigue Detection", B. Yin, X. Fan, and Y. Sun extend the idea of the local binary patterns based on Gabor features to account for the temporal aspect of human fatigue. In another paper, T. Germa, F. Lerasle, and T. Simon present "Video-Based Face Recognition and Tracking from a Robot Companion". The integrated still-to-video face recognition system proposed uses a number of features as persistent cues in a robust and probabilistically motivated way. // The special issue concludes with two papers on expression analysis and synthesis. I. Buciu and I. Nafornita present "Feature Extraction through Cross-Phase Congruency for Facial Expression Analysis". The authors use phase congruency maps to obtain discriminative features for facial expression analysis. In the paper entitled "Facial Expression Synthesis Based on Facial Component Model", L. Xiong, N. Zheng, S. Du, and J. Liu divide a face region into several partitions and construct local texture models to synthesize facial expressions.

This paper presents a novel algorithm for intermodulation products counting in multicarrier systems. Intermodulation products are caused by the nonlinearity of the amplifiers’ transfer characteristic. Along the transmission network, broadband amplifiers are needed to boost up the signal level and compensate the signal attenuation. When the number of carriers increases, the number of intermodulation products also increases extremely fast. In our approach, selective counting procedure by grouping intermodulation products was introduced. The presented solution shows considerable reduction in computational effort in counting the number of intermodulation products in the individual channel, especially if the number of products is very large.

It is a great pleasure welcoming all of you to the IEEE International Wireless Communications and Mobile Computing Conference (IEEE IWCMC 2015) in the beautiful Dubrovnik, Croatia! We are indeed delighted that this year's IEEE IWCMC lived up to its goal under the conference theme “Communications for the 21st Century, ” and continues its tradition of providing the premier forum for presentation of research results and experience reporting on the cutting edge research in the general areas of wireless communications and mobile computing. This year, we received more than 700 submissions from 42 countries. Each paper received at least three peer technical reviews, comprised of more than 450 TPC members from academia, government laboratories, and industries. After carefully examining all review reports, the IEEE IWCMC 2015 TPC finally selected about 36% high-quality papers for presentation at the conference and publication in the IEEE IWCMC 2015 proceedings. The conference program starts on Monday August 24th with a full day of Tutorials that are free of charge to all our attendees. Then, each day starts with a keynote speaker chosen from renowned world-class leaders in the area-Dr. Giuseppe Bianchi from University of Roma Tor Vergata, Italy, Dr. Mario Gerla from UCLA, and Dr. Slim Alouini from KAUST, highlighting the latest research trends in the wireless communications, mobile computing, and networks. This year, the technical sessions reflect the continued and growing interests in a wide range of spectrum, including wireless communications and networks, cross-layer design and optimization, mobile computing, wireless sensor networks, network security, and use of wireless technologies in social emergency applications. There are two special sessions composed of invited papers from renowned experts from around the world. Outstanding papers will be selected for four Special Issues in well known international journals. Our objective in the future is to reduce the acceptance rate further. A conference of this size cannot be organized without the hard work and dedication of many people. We want to thank all authors who submitted their manuscripts, as well as all the Symposia and Workshop Chairs/Co-Chairs, TPC members, invited reviewers for all their contribution to the conference. Special gratitude goes to IEEE as well as the IEEE Croatia Section for their technical sponsorship and to the University of Dubrovnik for the sponsorship. We hope you all enjoy IEEE IWCMC 2015 technical program and find this conference a productive opportunity to explore, exchange ideas, make new friends, and renew old ones and best of all come back next year to contribute in IWCMC 2016 that will be held in Cyprus.

The primary objective of this randomized phase III study was to show significant difference in median time to progression (TTP) in patients with advanced NSCLC treated with single-agent gemcitabine maintenance therapy versus best supportive care following gemcitabine plus cisplatin initial first-line therapy.Chemonaive patients with stage IIIB/IV NSCLC received gemcitabine 1,250 mg/m(2) (days 1 and 8) plus cisplatin 80 mg/m(2) (day 1) every 21 days. Patients achieving objective response or disease stabilization following initial gemcitabine plus cisplatin therapy were randomized (2:1 fashion) to receive maintenance gemcitabine (1,250 mg/m(2) on days 1 and 8 every 21 days) plus best supportive care (GEM arm), or best supportive care only (BSC arm).Between November 1999 and November 2002, we enrolled 352 patients (median age: 57 years; stage IV disease: 74%; Karnofsky performance status (KPS)80: 41%). Following initial therapy, 206 patients were randomized and treated with gemcitabine (138) or best supportive care (68). TTP throughout the study period was 6.6 and 5 months for GEM and BSC arms, respectively, while values for the maintenance period were 3.6 and 2.0 months (for p0.001 for both). Median overall survival (OS) throughout study was 13.0 months for GEM and 11.0 months for BSC arms (p = 0.195). The toxicity profile was mild, with neutropenia being most common grade 3/4 toxicities.Maintenance therapy with gemcitabine, following initial therapy with gemcitabine plus cisplatin, was feasible, and produced significantly longer TTP compared to best supportive care alone. Further studies are warranted to establish the place of maintenance chemotherapy in patients with advanced NSCLC.

Temporal comparison of lesions might improve classification between benign and malignant lesions in full-field digital mammograms (FFDM). The authors compare the use of volumetric features for lesion classification, which are computed from dense tissue thickness maps, to the use of mammographic lesion area. Use of dense tissue thickness maps for lesion characterization is advantageous, since it results in lesion features that are invariant to acquisition parameters.The dataset used in the analysis consisted of 60 temporal mammogram pairs comprising 120 mediolateral oblique or craniocaudal views with a total of 65 lesions, of which 41 were benign and 24 malignant. The authors analyzed the performance of four volumetric features, area, and four other commonly used features obtained from temporal mammogram pairs, current mammograms, and prior mammograms. The authors evaluated the individual performance of all features and of different feature sets. The authors used linear discriminant analysis with leave-one-out cross validation to classify different feature sets.Volumetric features from temporal mammogram pairs achieved the best individual performance, as measured by the area under the receiver operating characteristic curve (Az value). Volume change (Az = 0.88) achieved higher Az value than projected lesion area change (Az = 0.78) in the temporal comparison of lesions. Best performance was achieved with a set that consisted of a set of features extracted from the current exam combined with four volumetric features representing changes with respect to the prior mammogram (Az = 0.90). This was significantly better (p = 0.005) than the performance obtained using features from the current exam only (Az = 0.77).Volumetric features from temporal mammogram pairs combined with features from the single exam significantly improve discrimination of benign and malignant lesions in FFDM mammograms compared to using only single exam features. In the comparison with prior mammograms, use of volumetric change may lead to better performance than use of lesion area change.

Switching to digital technology from analog brought many changes, especially in storage of digital X-ray images. In cases of multiple fractures or need for imaging of a large area it is necessary to make more than one X-ray image for the observed part of body. This is often the case with spine, legs and arms. Radiologists use dedicated workstations to view X-ray results and need to manually switch between different images to be able to complete the examination of longer bone segments. Dedicated software gives some possibility of automatic or semiautomatic image stitching which allows combining of multiple images into one. In this paper we propose a method for automatic detection of bone position in X-ray images of arms and legs and their automatic alignment with one of image's borders.

Current standards for the compression of still and moving images use Discrete Cosine Transform (DCT) to remove spatial redundancy in images. Students specialising in image and video system engineering need to know why DCT is important in their field of interest and to understand the influence of DCT-based image compression on picture quality. Therefore, we have developed educational software, called DCTlab, that helps students to analyse DCT application in still image compression systems. This paper describes software characteristics, its application in a digital television course and learning outcomes.

Determining the breast density in mammograms is important both in diagnostic and computer-aided detection applications. Knowing the right breast density and having knowledge of changes in breast density could give a hint of a process which started to happen within a patient. Breast density could be rather easily estimated by dividing mammogram into fibroglandular and fat tissue. Mammograms suffer from a problem of overlapping tissue which results in possibility of inaccurate detection of tissue types. Fibroglandular tissue has rather high attenuation of X-rays and is visible as brighter in the resulting image. Small blood vessels and microcalcifications are shown as brighter objects with similar intensities as dense tissue. In this paper we try to divide dense and fat tissue by suppressing scattered structures which do not represent glandular or dense tissue in order to divide mammograms more accurately in two major tissue types. For suppressing blood vessels we have used Gabor filters of different size and orientation to detect edges of blood vessels and subtract them from the original image. Microcalcifications have been suppressed by combination of morphological operations on filtered image with enhanced contrast. Dense tissue has been segmented using different thresholds to avoid false detection.