Projects

FP7

FP7-IP Estimation and Control for Safe Wireless High Mobility Cooperative Industrial Systems (EC-SAFEMOBIL)

Project duration: 2011-2015

Funding: European Commission

Collaborating institutions: FADA-CATEC (Spain), DLR (Germany), Astrium (Germany), Indra (Spain), SELEX GALILEO (UK), Euroimpianti (Italy), University of Seville (Spain), University Duisburg-Essen (Germany) and University of Zagreb (Croatia)

CRV laboratory: LARICS

Project web page: http://www.ec-safemobil-project.eu/

The EC-SAFEMOBIL project is devoted to the development of sufficiently accurate motion estimation and control methods and technologies in order to reach levels of reliability and safety to facilitate unmanned vehicle deployment in a broad range of applications dealing with the landing of UAVs on mobile platforms in challenging conditions, the launching of an UAV from a manned vehicle, the surveillance, and the warehousing involving a large number of autonomous vehicles.

 

FP7-CSA Center of Research Excellence for Advanced Cooperative Systems (ACROSS)

Project duration: 2011-2015

Funding: European Commission

Collaborating institutions: University of Zagreb (Croatia)

CRV laboratories: AMOR, HOTLAB, LARICS, IPG

Project web page: http://across.fer.unizg.hr/across/project/project_info

The project Centre of Research Excellence for Advanced Cooperative Systems (ACROSS) aims at unlocking and strengthening the research potential of the Faculty of Electrical Engineering and Computing of the University of Zagreb (UNIZG-FER) in the area of cooperative systems related to robotics, networked embedded systems and renewable energy systems. Possible applications of such cooperative systems are numerous: advanced flexible manufacturing, renewable and sustainable energy generation, home and office automation, transport, logistics, environmental monitoring, healthcare, security and surveillance, human augmentation, etc.

 

FP7-SECURITY Toolbox Implementation for Removal of Anti-personnel Mines Sub-munitions and UXO (TIRAMISU)

Project duration: 2012-2015

Funding: European Commission

Collaborators: 23 partners from 11 European countries

The TIRAMISU project aims at providing the foundation for a global toolbox that will cover the main Mine Action activities, from the survey of large areas to the actual disposal of explosive hazards, including Mine Risk Education. The toolbox produced by the project will provide Mine Action actors with a large set of tools, grouped into thematic modules, which will help them to better perform their job. These tools will have been designed with the help of end-users and validated by them in mine affected countries.

 

COST

EURO-TELEPATH - Telepathology Network in Europe

Project duration: 2007-2011

Funding: European Commission; COST Agency (European Cooperation in the field of Scientific and Technical Research)

Collaborators: 18 EU countries and 3 non-COST institutions

CRV laboratory: VCL

Project web page: http://www.conganat.org/eurotelepath/

Coordination of research efforts to develop the most adequate technological framework for the management of multimedia electronic healthcare records (data and images) through the Internet. The Action will consolidate the most renowned research references in the field of informatics applied to Anatomic Pathology in order to eventually develop, with support of national and other European programs, the fusion standards to represent, interpret, browse and retrieve digital medical images while preserving their diagnostic quality as needed for clinical, learning and research purposes. In a later stage, this coordinated research shall bring about a comprehensive R&D project which will deliver a much needed world wide search engine based on WebServices. This will definitively open the path to integration, search, access, exchange and upgrade of digital pathological images and associated reports among different hospital information systems regardless of their location. Based on common standards developed by European Normalization Committee (CEN), the World Wide Web Consortium (W3C) and other bodies (DICOM, HL7, SNOMED), the direct result of this Action shall be a new Pathology Technical Framework (IHE Pathology) to be taken as a new reference standard by the specialized E-health industry as well as the entire medical community.

 

Other projects

Electronic differential for small electrical vehicle

Project duration: 2010-2012

Funding: Croatian Science Foundation

Collaborating institution: Exor, Zagreb (Croatia)

CRV laboratory: LARICS

Project web page: http://larics.rasip.fer.hr/eldiff/about_en.html

Due to the fact that design, construction and production of the entire electric car is unfeasible task for Croatian industrial factors because of current technology level in the country, the main goal of the project is to concentrate knowledge and experience of project participants on one particular segment of an electric city car, mainly, on design and construction of electronic differential prototype. Upon successful outcome of the project potential Croatian producers of the electronic differential (in the first place Exor as partner company) could compete on the international market with other producers of components for the future electric vehicles. Production of the electronic differential and competition for international market represents long-term objective of the project.

 

Exploration robot for firefighting units

Project duration: 2008-2010

Funding: National Foundation for Science, Croatia

Collaborating institution: Hrid, Zagreb (Croatia)

CRV laboratory: LARICS

Project web page: http://larics.rasip.fer.hr/projectInfo.php?id=26

The project resulted with the design and implementation of a prototype of a professional service mobile robot designed for execution of exploration and inspection tasks in dangerous environments. Built as a four flipper/track robot equipped with sensors (cameras, thermo-vision cameras, temperature sensors, explosive gas detectors etc.), its aim is to explore buildings, detect potential sources of danger (fire, explosive, poisonous gases, ruins, shaky obstacles etc.), and locate people caught in the accidents caused by flood, fire, earthquake or other natural and non-natural disasters. The use of a robot should minimize the risk of interventions of professional units (e.g. fire fighting, civil guard, police, military). The robot maintains wireless video and audio communication with the operator. On-board audio link enables conversation between the operator and the casualties. Having independently controlled tracks/flippers, the robot construction and maneuverability allows easy overtaking of obstacles, including climbing the steps.

 

Picture Quality Management in Digital Video Broadcasting

Project duration: 2007-ongoing

Funding: Ministry of Science, Education and Sports of the Republic of Croatia

Collaborating institutions: University of Osijek, University of Dubrovnik

CRV laboratory: VCL

The aim of the project is to propose an integrated picture quality management and evaluation in digital terrestrial television system. Digital terrestrial television has been studied in many parts of the world for the last decade or more. In Europe, proposed technology was agreed in a consortium known as the Digital Video Broadcasting (DVB) project and the system known as the DVB terrestrial specification (DVB-T) was standardized at ETSI. The standard defines a tool-kit of technical options but each implementation must select from these tools the precise set which matches the conditions in which the services will operate. The main DVB configurations that will be considered and tested in project are DVB-T and DVB-H (Handheld). DVB-H is new standard based on DVB-T but dedicated for handhelds and well suited to mobile communication. The main objective of the project is to specify, develop, integrate and validate in testbed and field trials a complete framework able to provide picture quality management in DVB-T/H systems. Project covers an entire DVB-T/H broadcast chain including source signals generation and coding, multiplexing techniques, distribution and reception. The aim is not to unify or impose a strategy on each individual entity of the chain, but to harmonize their functionality. Currently there are only partial or incomplete quality of service (QoS) definitions which do not adequately dimension picture quality from a user perspective. QoS parameters will be clearly identified, described and controlled all the way along the broadcast chain. To meet its objectives, the project will develop QoS concept using multi-dimensional QoS inputs and generating perceived picture quality as output. The picture quality measurement tool is an important element addressed by project in order to ensure that the end-user will benefit from end-to-end QoS provision. Expected results of the project include: picture quality parameters identification, generation of DVB-T/H simulation testbed, development of picture quality management model, setting up testbed for laboratory tests, design and implementation of perceived picture quality evaluation procedure. The results of the project will be verified by subjective assessment of picture quality for different sets of QoS parameters in laboratory conditions and by extended field trials on experimental DVB-T network. Project will provide results needed for the successful implementation of DVB-T/H systems in Croatia.

 

Intelligent Image Features Extraction in Knowledge Discovery Systems

Project duration: 2007-ongoing

Funding: Ministry of Science, Education and Sports of the Republic of Croatia

Collaborating institution: Ruđer Bošković Institute

CRV laboratory: VCL

Project web page: http://www.imagefeatures.org/

With the size of image databases increasing dramatically, the usefulness of such information is dependent on how well it can be accessed and searched and how well knowledge can be extracted from it. Our ability to generate data currently far outstrips our ability to explore, analyze and let alone understand it. Recently, there has been a growing and renewed interest in intelligent image management systems based on domain knowledge and applications. Although over the last decade or so some researchers have developed knowledge-based approaches, they rely almost exclusively on low-level features such as texture, color and intensity and have little high-level features interpretation capability. Also, they mainly operate in pixel domain, such that images (if compressed) need to be decompressed prior to any analysis or processing. Such an approach can slow down any application to the point of being impractical. This project will create intelligent methods for solving many difficult high-level image feature extraction and analysis problems, in which both local and global properties as well as spatial relations must be taken into consideration. All the techniques will be implemented to work in pixel and compressed domain (avoiding the inverse transform in decompression), thus speeding up the whole process. Hopefully, this research will lead to a better understanding of images and development of efficient methods for other image processing, pattern recognition and computer vision problems. The research results will have a wide range of useful applications, including but not restricted to human face image recognition, medical and biological image analysis, image and video compression, and content-based image indexing and retrieval. Initial investigation of intelligent image analysis and feature extraction techniques will be done, combining the principles from fields of multiresolution wavelet analysis, compression, computer vision, texture analysis and information retrieval. Next, by applying and extending those techniques we aim to solve a class of difficult image feature extraction problems. By combining various image analysis and signal processing techniques we hope to develop new high-level feature extraction methods, thus improving current state-of-the-art retrieval and classification methods. Using the resulting extracted features as a first step and input to data mining systems would lead to supreme knowledge discovery systems.

 

Monitoring the forest health status using remote sensing methods

Project duration: 2007-ongoing

Funding: Ministry of Science, Education and Sports, Croatian National Science Foundation

CRV laboratory: LRS

The purpose of the research is to establish and analyze the forest damage using the photointerpretation of color infrared (CIR) aerial photographs, for obtaining the data on the damage distribution, study the characteristics that have impacts upon forest health, and detect the possible agent of forest dieback. In addition, aerial multi-spectral surveying (digital camera MS-3100 and hyperspectral scanner V9) would enable mistletoe detection on fir trees of sufficient reliability that would support the research on the correlation between mistletoe and tree damage. A possibility of interpreting high-resolution satellite imagery (IKONOS) will also be investigated, in order to get quality data for the estimate and inventory of the forest damage degree, to be used in forest management, silviculture and protection.

 

Application of high resolution satellite imagery for management and planning in forestry

Project duration: 2011-2015

Funding: Hrvatske šume Ltd. Croatia

Collaborating institution: Hrvatske šume Ltd., Croatia

CRV laboratory: LRS

Main goal of this research project is to compare and retrieve the best method for interpretation of high resolution IKONOS satellite image, which would be simple and acceptable in practice and which would produce reliable data on stand parameters, applicable in forest management. Results of the research conducted in this project can be viewed through two basic aspects – comparison between visual and digital interpretation of satellite image and assessment of structural stand elements, based on the spectral reflection values in high spatial resolution IKONOS satellite image.

 

Autonomous Multiagent Automatic Assembly

Project duration: 2006-ongoing

Funding agency: Ministry of sciences, education and sport

CRV laboratory: LAMP

A complex system, composed of several robots performing assembly tasks is organized in a multiagent domain. Methods and procedures for coordination of the robots performing variety of real world actions are developed. Evolvability and reconfigurability procedures are developed and implemented to enable the system to adapt to the changing environment conditions based on sensory information fusion – 3D vision, laser etc., and to learn from the experience. The system is able to completely autonomously perform the most complex assembly tasks for two completely different products. It is moreover able to interact with humans, understand their actions and react accordingly so as to eliminate any possibility of injury on the humans entering the robots work space.

 

IGRAMO - Improving GRAsping Movements by predictions based on Observation

Project duration: 2009 - 2011

Funding agency: Ministry of sciences, education and sport

Collaborating institution: KTH Stockholm

CRV laboratory: LAMP

Project web page: http://www.sjever.fsb.hr/index.php?option=com_content&view=article&id=22&Itemid=15&lang=en

The main goal of the project was aiming at the design and development of an autonomous robotic object manipulation system. In this system, the artificial robotic agent is learning how to manipulate different objects based on observation of a human manipulating the same set of objects. This approach resembles the learning method used by children that observe their parents and imitate parent’s gestures. The motivation behind the work was the fact that future robots, in particular service robots, must be able to learn how to solve tasks in unknown environments in an easy and flexible way. These paradigms could be transposed to industrial robotic environments, enabling higher robustness and ease of work of such robotic systems.

 

Intelligent Cell for Feeding of Automated Assembly System

Project duration: 2006 - 2008

Funding agency: Croatian Institute of Technology

Collaborating institutions: EGO Elektrokontakt

CRV laboratory: LAMP

An intelligent cell composed of a Cartesian manipulator and dynamic vision system was developed for the purpose of assembly of a thermo regulator. The parts to be assembled are under no constraints regarding the position or lighting. Real time processing and fault free assembly was achieved.

 

Dual Arm Industrial Robot

Project duration: 2008 - 2010

Funding agency: Croatian Institute of Technology

Collaborating institutions: EGO Elektrokontakt, Končar niskonaponske skolpke d.d.

CRV laboratory: LAMP

To approach the level of flexibility of human hand, a setup based on two six DOF robots is proposed. The system is able to perform the most complex assembly tasks, performed solely by the human operators in the industrial shop floor. Original algorithms for visual perception and processing, on line path planning for dual arm robotic system, force/torque sensor etc. are developed and implemented.

 

Applications of Robots in Neurosurgery

Project duration (starting and ending year): 2010-ongoing

Funding agency: Croatian Institute of Technology

Collaborating institutions: Clinical Hospital Dubrava, Croatian Brain Research Institute

CRV laboratory: LAMP

To decrease the postoperative trauma, primarily in neurosurgical domain, we propose robotic assistant able to precisely locate area of interest in 3D space based on standard CT / MR input. Methods for registration between robot space and standard CT / MR spaces are developed. High speed force torque monitoring and reasoning procedures are developed and implemented among other.

 

Billing and Tracking System in Relation to Time Spent in Activity

Project duration: 2010-ongoing

Funding agency: Croatian Institute of Technology

Collaborating institution: Josip Juraj Strossmayer University of Osijek

CRV laboratory: LAMP

An autonomous tracking and billing system based on geo tagging and camera information fusion is developed. The system is able to localize objects of interests parallel, interpret the corresponding actions and deliver the receipt based on the activity analysis.

 

iRAP - EuroRAP – Road safety assessment programme

Project duration: 2009-ongoing

Funding: National program for road safety of Republic Croatia

Collaborating institution: Croatian Autoklub (HAK)

CRV laboratory: FTTS-CVG

iRAP's Star Ratings provide a simple and objective measure of the level of safety 'built in' to the road for car occupants, motorcyclists, bicyclists and pedestrians. Five-star roads are the safest and one-star roads are the least safe. Star Ratings are based on road inspection data collected through road inspection and analysis. Using specially equipped vehicles, video processing software and highly trained analysts, iRAP teams undertake detailed road inspections. These inspections focus on more than 30 different design features known to influence the likelihood of crashes as well as their severity. These features include intersection design, road cross-sections and markings, roadside hazards, footpaths and bicycle lanes

 

Mapping and Assessing the State of Traffic Infrastructure

Project duration: 2008-2011

Funding: Croatian science foundation, Institute of transport and communications, Zagreb, Croatia, Graz University of Technology

Collaborating institutions: University of Zagreb, Croatia; Institute of Transport and Communications, Zagreb, Croatia; Technical University Graz, Austria; Tilda d.o.o. Zagreb, Croatia

CRV laboratories: RUBIOSS, FTTS-CVG

Project web page: http://www.zemris.fer.hr/~ssegvic/mastif/index_en.shtml

This project addressed streamlining road safety inspection by relaxing the dependencies on trained human experts. In particular, we wished to find out whether reliable detection and recognition of different kinds of traffic signs and surface markings could be performed automatically by computer vision techniques. We focused on triangular warning traffic signs as the most frequent sign superclass along the Croatian local roads. We have shown that a performance of 100% recall (all warning signs correctly detected) with near 100% precision (very few false positive detections) and almost 100% correct recognition recognition rate is feasible.

 

Control of mobile robots and vehicles in unknown and dynamic environments

Project duration: 2007-2012

Funding agency: Croatian Ministry of Science and Technology

CRV laboratories: AMOR, FTTS-CVG

Project web page: http://act.rasip.fer.hr/projekt-opis.php?id=25

Mobile robotics is an interdisciplinary scientific discipline, which has been evolving very intensively for the last twenty years. A number of research groups of reputable universities around the world have developed their own mobile robot control systems, but a number of open-end research problems still remains. Particularly, the most control systems are tested in laboratory environments, which are assumed to be structured and stationary, but the direct, specific goal of the proposed project is to develop mobile robots and vehicles control system, which will enable fully autonomous navigation in unknown, unstructured and dynamic environments. Special research emphasis is on creating of a set of appropriate sensor array including perceptive (mono and stereo cameras, laser range finder, sonar sensor) and proprioceptive (gyro, compass) sensors. The most important expected result of the proposed research is a generic library of advanced mobile robot control functions, which can be easily applied to different designs of autonomous mobile robots and vehicles as well as to different applications in unknown and dynamic environments.

 

OptaGIS – GIS based transport infrastructure management system

Project duration: 2005-ongoing

Funding: Institute of transport and communications, Zagreb, Croatia; Faculty of Transport and Traffic Sciences, Zagreb, Croatia

Collaborating institutions: Institute of transport and communications, Zagreb, Croatia; Faculty of Transport and Traffic Sciences, University of Zagreb, Croatia; Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia

CRV laboratory: FTTS-CVG

A geo-information inventory for transport infrastructure has been actively developed at the Faculty of Transport and Traffic Sciences and the Institute of Transport and Communication from as a software product named OptaGIS. The developed system has been a successful tool in providing commercial road maintenance assessment service to local authorities all over Croatia since 2005. Compared to competing solutions, OptaGIS offers simpler assessment due to support of georeferenced video, better interoperability with other software (AutoCAD, etc.), as well as better representation capabilities due to support of different coordinate systems (Gauss-Krueger, WGS84, etc.).

 

System for the multisensor airborne reconnaissance and surveillance in the crisis and the protection of environment

Project duration: 2007-2008

Funding: Croatian Ministry of Science and Technology

Collaborating institutions: Faculty of Geodesy, University of Zagreb, Croatia; Croatian Mine Action Center - Center for testing, development and training, Zagreb, Croatia, Faculty of Transport and Traffic Sciences, Zagreb, Croatia, Pastor Group, Zagreb, Croatia

CRV laboratory: FTTS-CVG

In this project pre-commercial advanced civilian system for airborne multi-sensor reconnaissance and surveillance, composed of visible, near infrared, thermal, hyper spectral and navigational subsystems, has being integrated, tested, calibrated and demonstrated as a: system for airborne hyper-spectral reconnaissance and surveillance of oil spills sea pollution and airborne imaging system for decision support in uncertainty environment based on generic SMART (Space- and Airborne Mined Area Reduction Tool) methodology. Guidance and control system for active fire reconnaissance and surveillance using unmanned aerial vehicle has been realised and evaluation of public software for operational prediction of behavior of active fires and interpretation of fire scenes from thermal images are being performed.

 

Intelligent Control of Mobile Robots

Project duration: 2003 - 2006

Funding agency: Croatian Ministry of Science and Technology

CRV laboratories: AMOR, FTTS-CVG

Project web page: http://act.rasip.fer.hr/projekt-opis.php?id=4

In this project, methods, algorithms and structures for control of mobile robots are being investigated, which should ensure full autonomy of motion of mobile robots based on sensor measurements and without prior preparation of the environment. In order for the control system to fulfill this complex task, it must have the capabilities of estimating the mobile robot's current position and orientation, map building and path planning in the environment. When investigating solutions to these problems, core implementation is expected in intelligent methods, especially concerning neural networks and fuzzy logic, methods of stochastic reasoning and fusion methods of several sensors (mono and stereo cameras, laser range finder, sonar sensor, gyro, compass). The overall result is to be a generic library of programming functions of an intelligent control system applicable to a large class of various mobile robots, ranging from cheap domestic and leisure robots to expensive robots for complex applications in industry, construction, agriculture, etc.

 

Digitalization of the museum painting heritage

Project duration: 2006-2011

Funding: Ministry of science, education and sports

Museum painting collections are essential part of our culture heritage. The intention of this project is to approach systematically to digitalization of museum painting collections. The purpose of the research is to develop a model of digitalization which would enable the production of qualitative digital copies of originals. The research results will contribute to solving some specific problems which appear during the characterization of digital photographic systems for special purposes such as the demanding process of the accurate color reproduction of the art paintings.

 

Deployment of the Decision Support System for Mine Suspected Area Reduction

Project duration: 2008-2009

Funding: International Trust Fund for Demining and Mine Victims Assistance (ITF),

Collaborating institution: HCR Centre for testing, development and training Ltd.,

The project was aimed to support decision of the mine suspected area reduction in three communities having very different contamination, terrain, climate, mine laying history. Results of the application of Decision Support System for Mine Suspected Area Reduction in the Republic of Croatia in 2008 and 2009 are summarized as “Proposal to CROMAC decision makers to start process of the reduction of the MSA using evidences derived in the project” and “Recommendation to CROMAC decision makers to investigate areas outside of MSA in community Gospić, where were indicated hazardous parts”.

 

Deployment of the Decision Support System for Mine Suspected Area Reduction in Bosnia and Herzegovina

Project duration: 2009-2011

Funding: International Trust Fund for Demining and Mine Victims Assistance (ITF)

Collaborating institution: HCR Centre for testing, development and training Ltd.,

The Advanced Intelligence Decision Support System technology was applied from July 2009 to July 2011 on 171.12 km 2 near Bihać, Mostar and Trebinje in Bosnia and Herzegovina. The goal was to reliably determine status on 40.2 km2 of the mine suspected area (MSA). The key sources of the new data are aerial multisensor images acquired in November 2010 and satellite images Ikonos 2 and WorldView 1.

 

Autonomy-Oriented Computer Structures

Project duration: 2006-ongoing

Funding: Ministry of Science, Education and Sports of the Republic of Croatia

Collaborating institutions: Faculty of Engineering, University of Rijeka; Faculty of Organization and Informatics, Varaždin

CRV laboratory: RUBIOSS

Project web page: http://zprojekti.mzos.hr/page.aspx?pid=97&lid=2

Natural and artificial worlds are filled with multi-entity systems with the following features: Autonomous – entities are rational individuals that act independently; Emergent – they exhibit complex behaviors that are not predefined in the behavior of individual entities; Adaptive – They change their behaviors in response to changes in the environment in which they are situated; Self-organized – They are able to organize the elements to achieve the above behaviors. Autonomy-oriented computing (AOC) is a new computing paradigm (J. Liu et al. 2002; J. Liu et al. 2005; M. Dorigo, 1996) for solving complex problems and modeling of complex systems. Theory of AOC is in its early stages of development and is based on theory of multiagent systems (J. Ferber, 2001; G. Weiss, 1999; M. Wooldridge, 1996; 2002;). Extended theory, obtained by including modal operators for the agent blindness and forgetfulness, extending the entities by adding biological and social properties like, for instance: birth, inheritance, dying, fuzziness in mutual communication and cooperation, will be applied for complex system modeling in the areas of computer science (multimodal biometric systems, computer vision systems, knowledge representation) and information science (knowledge management systems, spoken dialog systems). The expected results of this multidisciplinary program are: development of the theory and models of the AOC structures and multiagent systems, development of AOC methodology and multiagent modeling. We will evaluate the research results on modeling of knowledge representation systems, spoken dialog systems, multisensor multimodal biometric systems and computer vision systems. Participants in the program are Faculty of Electrical Engineering and Computing, Faculty of Organization and Informatics, Varaždin, University of Zagreb and Faculty of Philosophy, University of Rijeka.

 

Model based traffic sign detection

Project duration: 2010-2011

Funding: Ministry of Science, Education and Sport of the Republic of Croatia

Collaborating institution: Graz University of Technology

CRV laboratory: RUBIOSS

Project web page: http://projects.unizg.hr/en/projects?@=5try#proj_10655

The project aims at providing added value to the existing semi-industrial project by: (i) development of novel models for traffic sign detection and recognition; (ii) strengthening the existing cooperation between the two computer vision research groups with the goal to facilitate future joint FP7 project proposals; and (iii) supervision and training of a young Croatian researcher by a distinguished Austrian professor.

 

Advanced techniques for detection of traffic control devices

Project duration: 2012.03.01-2012.09.01

Funding: University of Zagreb

Collaborating institution: Graz University of Technology

CRV laboratory: RUBIOSS

This project addresses several advanced techniques for detection of traffic signs and road surface markings .The main line of the project investigates occluding boundaries as a novel cue for generic object detection. Additionally, we shall investigate the use of stereo in order to support the detection procedures by the estimated dynamics of the camera pose. Finally, we shall also address simultaneous detection of traffic signs from heterogeneous classes.