Navigation : EXPO21XX > VISION 21XX > H15: Research and Universities > University of Zagreb
  • Offer Profile
  • Research topics of interest include:
    • Image processing and analysis
    • Biomedical image analysis
    • Computer vision for automotive and traffic/transportation applications
    • 3D object reconstruction
    • Passive and active stereo vision and visualization techniques
    • Virtual reality
Product Portfolio
  • Center of Excellence for Computer Vision

  • The University of Zagreb has approved the project "Center for Computer Vision" that is funded through the University development fund.
    The project coordinator is prof. Sven Loncaric. Members of the Center are professors, associate and assistant professors and doctoral students spanning four faculty departments (ZESOI, ZEMRIS, ZARI, ZRK) and six other constitutional units of the University (Faculty of Mechanical Engineering and Naval Architecture, Faculty of Transport and Traffic Engineering, Faculty of Graphic Arts, Faculty of Kinesiology, Faculty of Forestry and Faculty of Geodesy).
    The main goal of the project is to establish an internationally recognized excellence center to better integrate researchers in the area of computer vision from various constitutional units of the University. The most important project tasks are aimed toward establishing the organization of the Center, toward promoting the Center, toward providing support to Center members when establishing contact with another similar centers in the world, toward achieving stronger cooperation with the industry and enabling transfer of technology, and toward establishing research seminars and annual Center workshop.
      • Retinal Image Analysis of Diabetic Patients

      • The project addresses a wide-spread medical condition called diabetic retinopathy. The cause of diabetic retinopathy is damage to blood vessels of the retina. Diabetic retinopathy is the leading cause of blindness. Diabetic patients with type 1 and type 2 diabetes have a greater risk for this condition. Suffering from severe diabetes for a longer period of time increases the chance of getting retinopathy.
        The goal of the project is to develop image processing and analysis methodology for computer-based interpretation of images obtained by retinal photography. The goal of quantitative image analysis is to detect if a patient retina photograph shows a normal retina or a retina affected by diabetic retinopathy. This goal will be achieved by recognizing the relevant structures on the retina including blood vessel tree, fovea, and blind spot. In the second phase of the project, the goal is to quantify the progress of the disease. 
      • Automated Visual Inspection

      • Visual inspection and quality control is a required component of any modern production line. For high-volume production visual inspection must be fully automated and integrated into the production line.
        The aim of this project is to develop customized image processing algorithms and complete visual inspection systems that can be used to improve the manufacturing process in the Elektro-kontakt's manufacturing plant.
      • Self-balancing Vehicle with Single Wheel

      • The aim of the project was to develop personal self-balancing single-wheel vehicle, which fulfills these requirements: small dimensions, mass and energy consumption, with the possibility of charging the battery from a regular network port or a renewable source. Regulatory and adaptational algorithms and the digital signal processing allow the drive of the vehicle with no learning period.The vehicle consists of these parts:
        1. lithium-ferro phosphate batteries
        2. sensor system for tilt detection implemented as an electronic printed circuit board with accelerometers and
        3. processor - the central computer system which performs all necessary processing of the data received from
        4. wheel with an electric motor located in its center
      • panoVRama - A System for Multi-Projector Tiled Visualizations

      • Multidimensional data sets are important in many areas of human activity such as image and volume sequence acquisition in biomedicine, industrial process control, remote sensing in meteorology and agriculture. 3D visualization provides means for easier interpretation, analysis, understanding of complex phenomena, and is basis for virtual reality simulations.
        The goal of the project is realization of tiled visualization system using multiple projectors. The system consists of nine tiled projectors that provide a better impression of immersion in the virtual environment. Several important problems have been researched in the project such as the problem of brightness correction in the overlap areas of two or more projectors, correction of varying color balance between projectors, geometrical camera calibration problem for multiple projectors, and synchronization of multiple computers. A distributed Linux-based client-server system has been developed to distribute panoramic image and video content to multiple video projectors.
      • Real-Time Image Processing for Automotive Panoramic Visualization

      • The project addresses emerging requirements of automotive industry for advanced driver assistance systems (ASAD). This system enables the driver to see the complete perimeter around the vehicle and serves as parking assistance for standard vehicles and special purpose vehicles. The system consists of multiple cameras mounted on the vehicle and custom FPGA based processing unit for real time image acquisition, processing and visualization. The cameras are equipped with fish-eye lens with field of view of more than 180-degrees. The image from each camera undergoes lens distortion correction, perspective correction, and stitching to create a view from a viewing point placed above the vehicle.
        The second goal of the project is to develop real time image processing and computer vision methods suitable for implementation on FPGA hardware platform and software for automatic camera calibration and multiple camera system calibration. Significant FPGA platform restrictions and limitations impose great challenges on implementation of methods for image processing.
        The third goal of the project is to develop a methodology and software for dynamic calibration of the cameras, which is required for adaptive changes in camera calibration parameters due to variable vehicle load. The method will be implemented using FPGA hardware and will help achieve a high quality visualization of the vehicle surroundings regardless of the number of persons in the vehicle or the load. The fourth goal is an alternative type of visualization obtained by projection of the ground plane onto a bowl-shaped surface. The resulting visualization has the advantage that the distant portions of the scene are displayed in such a way that scene details are better presented.
      • Real-Time Guidewire Tracking for Intravascular Interventions using C-Arm Imaging Device

      • Visual tracking of moving objects is one of important open problems in the field of computer vision. Due to problem complexity instead of general purpose only solutions for well-defined applications are considered. Visual tracking in the field of bio-medicine is often based on transmissive or projective acquisition models making existing procedures not applicable due to incompatibility of the usual acquisition model. Better results can be achieved by utilizing methods specifically adapted to the acquisition model.
        Minimally invasive endovascular interventions are good examples of surgical procedures that would be impossible without proper imaging equipment. Such interventions are the preferred treatment method for various vascular diseases. During intervention surgical instruments, such as guidewires and guiding catheters, are introduced into the vascular system and must be navigated to a point of interest, usually a pathology. The task of the medical imaging chain is to provide the surgeon with the best possible information required for successful navigation.
        The aim of the project was to develop a real-time capable system that tracks the guidewire in 2D image sequence and back-projects the found guidewire into 3D space.
      • Automated Visual Inspection of Plastic Products

      • The main task of machine vision systems is providing computer understandable descriptions of objects from either single image or whole array of images. One of such problems is encountered in automated visual product inspection. An automated visual inspection system must discover and classify possible defects from product images and should be fairly quick and robust. These requirements are needed if automated systems are to replace human inspection which has many drawbacks, mainly caused by tiredness and slowness (modern production methods usually facilitate productions speeds humans can't cope with). The main aim of this research project is developement of algorithms for defect detection and classification along with the development of the complete prototype system.
      • Functional X-ray Neuro Image Analysis

      • Quantitative functional measurements can significantly improve diagnostic value of x-ray angiograms. Of several different functional parameters three have clinical importance: blood flow velocity, perfusion and diffusion. X-ray imaging can provide only first two parameters, of which perfusion imaging is more important due to several important hemodynamic parameters: cerebral blood volume (CBV), cerebral blood flow (CBF) and mean transit time (MTT).
        The primary aim of this project was to exploit good characteristics of x-ray imaging and develop functional analysis techniques that will help in studying of the human physiologic characteristics associated with known pathologies.
      • 3D Quantification of Intracerebral Brain Hemorrhage

      • Computed tomography (CT) allows three-dimensional (3-D) anatomical imaging of brain abnormalities such as human spontaneous intracerebral brain hemorrhage (ICH). With computerized image analysis, it is feasible to characterize the pathology of a selected volume of interest. The proposed research focuses on 3-D quantitative analysis to study the early evolution of the ICH. The underlying hypothesis is that the ICH volume and structure is related to the mortality and morbidity. Patients having ICH are scanned four times: within three hours after first symptoms, one hour later, eight hours later, and within twenty hours after first sympt oms. During the course of the illness, 3-D changes in ICH volume and structure can be observed and analyzed. The important ICH features are volume, position in space, and shape of primary and edema region. The preliminary studies indicate that the ICH volume is significant for the survival of the patient. The position in space must be measured with respect to an invariant 3-D coordinate system so that the movement of the ICH accross the scans can be determined. To achieve invariance it is necessary to perform registration of brain CT images from two scans. We have recently developed an Iterative Principal Axes Registration (IPAR) algorithm to register 3-D multi-modality brain images. We have also developed 3-D spatially weighted region growing algorithms with adaptive clustering for segmentation of ICH regions. Finally, shape features will be computed to correlate the shape evolution of the ICH to mortality and morbidity. In addition, the characteristic behavior of ICH can be correlated with the patient response to the medical treatment with the purpose of evaluating the treatment earlier during the course of the illness. It is expected that the proposed research would provide a computerized system for analysis of the ICH through the characteristic changes in ICH volume and structure during the course of the illness.
      • Medical Teleconsulting

      • Providing health care services in rural areas has shown to be a challenging task. Low concentration of population makes health care institutions not self-sustainable so typically only small ambulances with general practitioners are available in rural areas. In such situation, one of the biggest problems is lack of medical specialists.
        Significant improvement in providing health care services in mentioned areas can be achieved by use of electronic means of communication. Rapid Internet development in recent years has made possible new medical applications including telemedicine. Telemedicine helps connect rural areas to large medical centers through exchange of medical information between distant locations.
        The Virtual Polyclinic, a web-based system provides tools to exchange medical information between general practitioners in rural areas and specialists in large medical centers.

      • Cardiac Doppler Ultrasound Image Processing and Analysis

      • Analysis of cardiac ultrasound images is an important tool for diagnosis of cardiac diseases. The hypothesis of the research project is that Doppler ultrasound images of aortic outflow velocity profiles have different features for healthy and diseased cases. The goal of the project is to develop a signal and image analysis methodology to extract relevant structures and features from DICOM Doppler ultrasound images to be used for statistical analysis of aortic outflow profiles for a larger number of patients.

      • EEG Signal Analysis of PTSP Patients

      • Analysis of Electroencephalography (EEG) signals is a standard procedure in for detection of many sorts of problems associated with brain function. EEG measures spatially located electrical activity of the brain by measuring differences of electrical potentials between electrodes located on the head. The graphical representation of electrical potential differences over time is called Electroencephalogram, and physicians use these signals or wave patterns to detect various disorders. Any detectable disorder will manifest as a deviation from the normal wave patterns, and certain medical conditions produce different types of deviations. Alongside visual evaluation of EEG signals, computer based digital signal analysis and processing techniques can also be used. Such techniques provide means for automatic, fast and reliable detection of abnormal EEG patterns, and their quantitative analysis.
        The overall objective of the research project is discrimination between aggressive and non-aggressive variants of PTSP disorder. We are trying to find significant features in EEG signals that can be used to discriminate two PTSP variants and to construct a detection and discrimination algorithm. The research efforts are directed toward characteristic wave pattern analysis in time frequency domain. The primary focus is on the use of wavelet transform and its different variants. Evaluation and testing of the methods to be developed will be performed on real patient EEG signals.
      • Intelligent Methods for Image Processing and Analysis

      • Image analysis and scene interpretation are complex tasks that require knowledge about objects contained in the scene and about their mutual relationships. Conventional image analysis approaches are based on relatively simple techniques that only utilize low-level information obtained from pixel intensity values. The major limitation of these approaches is lack of high-level knowledge.
        The goal of the project is to develop intelligent knowledge-based methods for object detection, recognition, and tracking that are robust and accurate. The proposed approach utilizes both low- and high-level knowledge for scene interpretation. To achieve this goal we will employ knowledge-based techniques such as 3-D model-based methods, neural networks, evolutionary algorithms, expert systems, and intelligent agents and apply developed techniques to the problems of medical, face, and range image analysis.
      • Detection of Vegetation for Traffic Applications

      • The goal of the research project was to develop an image analysis method for detection and recognition of vegetation alongside railroad tracks. Images for analysis are obtained from a video camera monitoring the immediate surrounding of the railroad tracks, mounted on the train. The detection of vegetation is based on the color information while recognition of different types of vegetation is based on texture features. The goal of detection is to identify image regions covered with vegetation. The recognition has to discriminate narrow-leaved and wide-leaved vegetation in the detected areas. The method has been designed to operate in real time in good lighting conditions.
      • Virtual Physiological Human Network of Excellence

      • The VPH Network of Excellence (VPH NoE) is designed to foster, harmonise and integrate pan-European research in the field of:
        1. patient-specific computer models for personalised and predictive healthcare, and
        2. CT-based tools for modelling and simulation of human physiology and disease-related processes.
      • Face Image Analysis for Biometric Applications

      • Modern biometric passports contain different biometric data, but they all contain ID photographs. Such photographs are intended for automatic face recognition and validation which require image quality sufficient to support recognition methods. The quality level is estimated by checking a number of predefined requirements that ID images have to satisfy (defined by ICAO standard). Besides manually by human operator, such testing of images can be done automatically to some extent. Simpler requirements can easily be done automatically, but the more complex tests require knowledge of the scene composition. Once appropriate regions of interest in ID photo are identified it is easy to check more complex quality requirements. Examples of complex requirements are: no hair over eyes, face dimensions compared to image dimensions, no red eyes etc. The difficult part is to to find objects of interest in ID photographs.
        The goal of the project was to segment input ID image into five most common regions. The most significant problem came from the fact that segmentation was a perquisite for quality check and hence hat to perform well on poor quality images (poor quality would be detected subsequently). The difficult problem required a substantial use of prior knowledge of the problem. Several knowledge based approaches have been tried: expert systems, neural networks, boosted classifiers, graph cuts etc. for training of learning based approaches, a image database has been formed with careful selection of images to include all possible variations in ID images. The final result of the project was a prototype software application.
      • COST Action B21 - Physiological Modelling of MR Image Formation

      • Since its introduction in the mid eighties, the growth in the use of magnetic resonance imaging (MRI) in medicine has been spectacular to the present position where it is widely used throughout the USA, Europe and Japan as a primary diagnostic imaging modality. Indeed, Young (1990) has hailed MRI as 'possibly the most powerful in-vivo diagnostic tool yet discovered' with 'the single most exciting thing about it being its scope'. There is considerable opportunity for further advances in utilisation and efficacy through making full use of all the quantitative image information available in MRI experiments and this Action seeks to take advantage of these possibilities, drawing together this data in order to interpret tissue structure and physiology in a totally original way as described on page 3.
        Current biomedical research (like drug development, genomics or the investigation of the mechanism of many natural and pathological processes such as ageing, cancer or multiple sclerosis), and clinical diagnosis depend more and more strongly on the availability of in-vivo information about local morphology or physiological processes on the cellular level. During the last few years, the development of cellular imaging techniques has introduced the possibility of overcoming some of the underlying problems and establishing new innovative ways of gaining insight into the functioning of living organisms.
        The Action will draw on the significant progress made during COST Action B11 ‘Quantitation of Magnetic Resonance Imaging Texture’, extending and developing the work into the fundamental issue of linking NMR and MRI measurements and other complementary image information with tissue structure and function. It also seeks to take full advantage of the effectiveness of the exceptionally strong working partnership developed during B11.