Menu Content/Inhalt
Pattern Recognition Group (PR-CVC) PDF Print E-mail

Comprised of researchers from the Computer Vision Center (CVC), Autonomous University of Barcelona (UAB), University of Barcelona (UB) and Open University of Catalunya (UOC). The main areas of research are:  

Colour and Texture

The CVC colour group focus its research in areas related to computational colour within computer vision. Our long term objective is to create computer algorithms that simulate human perception and categorisation of colour. To achieve this aim, we study colour as a visual cue in its context. The colour group is comprised of researchers (6 Ph. D. and 11 Ph. D. Students). Our main research lines are: Colour constancy,Colour sharpening,Colour-Texture perception,Colour Naming,Texture description,Colour & Texture grouping.


Pattern Recognition and Document Analysis

Document Analysis is a discipline that combines image analysis and pattern recognition techniques to process and extract information from documents from different sources. Sources include either raster formats, after scanning paper-based documents, or electronic formats such as ps, html, pdf, etc. Document Analysis consists of three major research subfields: paper layout analysis, optical character recognition and graphics recognition. The Document Analysis Group of the CVC has research and development experience in the following concerns: symbol recognition, indexing and browsing by graphical content, sketchy interfaces, diagrammatic reasoning and visual languages for graphic documents, graphics recognition architectures, reading systems for forms and structured documents, camera-based OCR, fingerprint recognition. More information about the research members of this area can be found in the DAG web site.


Medical Imaging Laboratory

Today Medical Imaging expands beyond the simple visualization and inspection of anatomic structures. Involving a large set of image modalities medical imaging provides meaningful structural, anatomical and functional information about human organs that serves to guide intervention procedures, plan surgeries and help the follow-up of patient diseases. Key issues of robust medical imaging analysis are represented by the following Computer Vision techniques:

a) Deformable models with their profound roots in estimation theory, optimization, and physics-based dynamical systems, represent a powerful approach to the general problem of image segmentation, image registration and 3D reconstruction.
b) Advanced classification techniques as adaboost, co-training, editing, etc. provide robust tools to learn and extract image features for straightforward tissue characterization, as well as combined with deformable models allow to capture and describe the variability of biological shapes in medical images.

Object Recognition

The human visual system can recognize unprimed views of common objects at sustained rates in excess of 10 per second (between 5.000 and 30.000 different objects!). The reproduction of the visual object recognition in the machine has been shown to be far more complex than the initial predictions made by researchers in the early 80s. A common assumption is that this process is based on a feedforward feature extraction hierarchy. From a computational perspective this assumption leads to several interesting questions: How to select the best set of features? Can these features be learned from examples? Which kind of classifiers are suited for this kind of architectures? How to build robust models from local features? The Object Recognition group is working along these lines to develop robust recognition systems to be used in non controlled environments.

Graphic Visualization and Modelling

The main research lines are focused on developing mathematic tools, interactive visualization environments, and augmented reality interfaces (especially) designed for biomedical imaging analysis. The modeling phase comprises both the geometric extraction (outline segmentation and shape description) and the dynamics (movement tracking and deformations) of the organ or cell under study. The visualization phase is focused on the development of virtual environments that allow an easy manipulation of the computational model, locally or remotely.

PR-CVC Members