If you have any questions regarding possible cooperation, please contact the research team management.

The team offers end-to-end development of turnkey computer vision applications for the industry. In most cases, solutions for optical control of product quality, correct assembly verification and optical recognition of different components or properties of products, parts or semi-finished products at different production stages are developed. These applications mostly improve production efficiency (faster processes, acceleration of machine cycle by replacing time-consuming manual control by an automatic system), quality (prevention of defective work by forbidding the start of assembly, timely detection of defects), and provide for simple data recording during production; these data can be used later for various analyses.

System design typically includes:

• Specification of the requested function
• Thorough analysis of conditions
• Analysis of technical properties in terms of optical information processing
• Analysis of human visual information processing
• Design of algorithms for processing, debugging and optimisation of algorithms
• Software development and implementation
• Hardware logistics, installation, software installation, etc.
• Commissioning

Experience confirms that optical control based on a camera system and evaluation software can often improve efficiency and quality of production. The same applies to similar segments, such as the service industry, agriculture, etc., generating economic benefits for the client. Secure remote administration and service staff training can be provided, if needed.

See also:

• CEDA – Central European Data Agency, a.s. –  a system for detection of traffic objects on roads developed in a project funded by the Technology Agency of the Czech Republic  
• Novem Cars Interior Design GmbH, Vorbach, Germany – a system for optical measurement of paint waviness; contract-based research
• Robotic system for removal of dangerous substances (Certicon, Asekol)
• Analysis of various plant parameters (Photon Systems Instruments, VÚGZ AVČR)

Our team offers an end-to-end development of software tools for medical image analysis. Modality being used imposes no restrictions; the most frequent imaging modalities are CT, MRI and USG. Most frequently developed applications are those for automatic or semi-automatic organ identification and highlighting or measuring of organ parameters or pathologies in recorded images. Tasks being solved include identification of various pathologies, calculations of the circumference, volume or size, vascular tree monitoring, etc.

System design typically includes:

• Specification of the requested function
• Thorough analysis of conditions, type and properties of image data
• Analysis of the process of visual information processing by a physician
• Design of algorithms for processing, debugging and optimisation of algorithms
• Software development and implementation
• Hardware logistics, connectivity, software installation, etc.
• Commissioning

Experience confirms that application of computer vision methods often improves the efficiency of physician’s work by providing (semi)automatic parameter calculations that would otherwise require a lot of physician’s time, or by pre-processing of large data volumes and identification of patients with more serious conditions, who can then receive more of the physician’s attention. Service staff training can also be provided as necessary.

See also:

• Segmentation of liver parenchyma, identification of hepatic vascular tree (University Hospital in Pilsen)
• Analysis of angiographic examination results (Nemocnice Na Homolce Hospital, Prague)

The team provides creation of three-dimensional models or design of 3D scanning systems for applications in industry, health industry, biology, agriculture, man-to-machine communication systems and other areas. Typical tasks include a 3D model of a human head or a human body to create a virtual 3D model (an avatar) that can be used in man-to-machine communication systems for audiovisual speech synthesis of a natural language or a sign language, in computer game industry or in other systems based on virtual 3D models of a human body. Another typical task is to create a 3D model of a component, product, plant, patient or body organ, using either 3D data available (e.g. in the health industry), or 3D surface scanning method. Yet another task is to design 3D scanning systems where more than one model is required and a system for repeated creation of 3D models must be developed.

Development of a 3D model typically includes:

• Identification and analysis of the scanned object
• Thorough analysis of conditions that can occur during scanning
• Design of a suitable process, 3D data capturing and pre-processing/analysis of these data
• The actual model development and its export into the required format

Design of a 3D scanning model typically includes:

• Identification and analysis of the scanned object
• Thorough analysis of conditions that can occur during scanning
• Selection of a suitable technology and proposal of a suitable process
• Hardware logistics, development of software to capture, edit or post-process 3D scans
• Commissioning

Experience confirms that 3D modelling is an efficient way to create models for different applications – from models of manufactured components used to control production on one hand to rapid prototyping on the other, to models of human heads for the “virtual counterpart” system, and medical 3D models of body organs or pathologies facilitating efficient image data analysis by physicians.

See also:

• 3D model of an avatar for the sign-language dictionary (University of West Bohemia, Masaryk University, UJP Praha a.s.)
• Design of a system for 3D plant modelling (Photon Systems Instruments (PSI), VÚGZ The Czech Academy of Sciences)
• Virtual Moderator – Česká hlava Endowment Fund, Scientific Café Olomouc, 10 October 2006 (3D scan and animation of a virtual model of a human head of the moderator of the evening – Mr. Václav Marek of the Česká hlava Endowment Fund)

The team provides motion capture analysis (MOCAP) – using the VICON optical system with passive reflective markers that can be placed on rigid objects, such as production line robots or quadrocopters, as well as human body using a special suit on which the markers can be applied easily. Motion capture data can be used to analyze movements in the health industry (physiotherapy, orthopaedics, etc.) or sports, as well as to provide realistic animations of body movements to design man-to-machine communication systems (sign-language synthesis), applications for computer game or filmmaking industries. The system also includes a unique CARA headset enabling efficient recording of accurate facial animation data. The pertinent software, that is a part of the system, is used to process different types of tasks (animations, medical applications, rigid objects, etc.).

Recording of 3D motion data analysis typically contains:

• Identification and analysis of the scanned object
• Thorough analysis of conditions that can occur during scanning
• Design of a suitable process, recording, pre-processing, analysis and post-processing of 3D motion data
• Delivery of data or processing results/analyses in the required format

Experience confirms that the use of motion-capture system data is an efficient way to create realistic movement for body and facial animations, obtain analyses of human body movements for medical applications and to provide basic information on the movement of tracked (e.g. robotic) models. 

See also:

• Virtual moderator – Česká hlava Endowment Fund, Scientific Café Olomouc, 10 October 2006 (3D scan and animation of a virtual model of a human head of the moderator of the evening – Mr. Václav Marek of the Česká hlava Endowment Fund
• Virtual moderator – virtual counterpart of J. A. Komenský, University of West Bohemia – official opening of the new building of the former Art and Design Institute, 18^th October 2012 (3D scan and animation of a virtual model of the head of J. A. Komenský based on his bust)
• Virtual moderator – virtual counterpart of Mona Lisa, University of West Bohemia – gala evening before launching the Pilsen – European Capital of Culture 2015 event, building of Ladislav Sutnar Faculty of Design and Art, 16^th January 2015 (3D scan and animation of a virtual model of the head of Mona Lisa based on a plaster bust created according to the Mona Lisa painting)
• Facial 3D animation of a small insect model for a film prepared by Studio Pokrok, a joint project