Automated Mapping System Research Articles

Guiding ablation strategies for ventricular tachycardia in patients with structural heart disease by analyzing links and conversion patterns of traceable abnormal late potential zone.

Substrate-based ablation can treat uninducible or hemodynamically instability scar-related ventricular tachycardia (VT). However, whether a correlation exists between the critical VT isthmus and late activation zone (LAZ) during sinus rhythm (SR) is unknown. To demonstrate the structural and functional properties of abnormal substrates and analyze the link between the VT circuit and abnormal activity during SR. Thirty-six patients with scar-related VT (age, 50.0 ± 13.7 years and 86.1% men) who underwent VT ablation were reviewed. The automatic rhythmia ultrahigh resolution mapping system was used for electroanatomic substrate mapping. The clinical characteristics and mapping findings, particularly the LAZ characteristics during SR and VT, were analyzed. To determine the association between the LAZ during the SR and VT circuits, the LAZ was defined as five activation patterns: entrance, exit, core, blind alley, and conduction barrier. Forty-five VTs were induced in 36 patients, 91.1% of which were monomorphic. The LAZ of all patients was mapped during the SR and VT circuits, and the consistency of the anatomical locations of the LAZ and VT circuits was analyzed. Using the ultrahigh resolution mapping system, interconversion patterns, including the bridge, T, puzzle, maze, and multilayer types, were identified. VT ablation enabled precise ablation of abnormal late potential conduction channels. Five interconversion patterns of the LAZ during the SR and VT circuits were summarized. These findings may help formulate more precise substrate-based ablation strategies for scar-related VT and shorter procedure times.

Computational Intelligence Based Point of Interest Detection by Video Surveillance Implementations

Latest advancement of the computer vision literature and Convolutional Neural Networks (CNN) reveal many opportunities that are being actively used in various research areas. One of the most important examples for these areas is autonomous vehicles and mapping systems. Point of interest detection is a rising field within autonomous video tracking and autonomous mapping systems. Within the last few years, the number of implementations and research papers started rising due to the advancements in the new deep learning systems. In this paper, our aim is to survey the existing studies implemented on point of interest detection systems that focus on objects on the road (like lanes, road marks), or objects on the roadside (like road signs, restaurants or temporary establishments) so that they can be used for autonomous vehicles and automatic mapping systems. Meanwhile, the roadside point of interest detection problem has been addressed from a transportation industry perspective. At the same time, a deep learning based point of interest detection model based on roadside gas station identification will be introduced as proof of the anticipated concept. Instead of using an internet connection for point of interest retrieval, the proposed model has the capability to work offline for more robustness. A variety of models have been analysed and their detection speed and accuracy performances are compared. Our preliminary results show that it is possible to develop a model achieving a satisfactory real-time performance that can be embedded into autonomous cars such that streaming video analysis and point of interest detection might be achievable in actual utilisation for future implementations.

Agile TMS: A multi-locus system for rapid and automatic spatial targeting and mapping

Agile TMS: A multi-locus system for rapid and automatic spatial targeting and mapping

MappyLand: Fast, Accurate Mapping for Console Games

We present MappyLand, a rewrite and enhancement of the earlier Mappy automatic game mapping system, which leverages instrumentation of a game console emulator to produce, from a sequence of game inputs, accurate annotations for action/adventure games including object detection and tracking, in-game camera movement, grid-based tile maps, and links between identified disparate spaces. The overhead of generating these annotations is on the order of one millisecond per observed frame. We also show a higher latency (but still online) algorithm for merging together previous observations of distinct game maps for the purposes of agent localization across a long period of time. Specifically, our system can determine a consistent graph of game rooms from a set of strings of game rooms, capturing behaviors like backtracking and synthesizing observations from multiple play sessions. Altogether, this fast, accurate approach to mapping yields new and useful knowledge representations and expedient ways to produce new datasets given just a game and some example play.

A digital mapping application for quantifying and displaying air temperatures at high spatiotemporal resolutions in near real-time across Australia.

Surface air temperature (Ta) required for real-time environmental modelling applications should be spatially quantified to capture the nuances of local-scale climates. This study created near real-time air temperature maps at a high spatial resolution across Australia. This mapping is achieved using the thin plate spline interpolation in concert with a digital elevation model and ‘live’ recordings garnered from 534 telemetered Australian Bureau of Meteorology automatic weather station (AWS) sites. The interpolation was assessed using cross-validation analysis in a 1-year period using 30-min interval observation. This was then applied to a fully automated mapping system—based in the R programming language—to produce near real-time maps at sub-hourly intervals. The cross-validation analysis revealed broad similarities across the seasons with mean-absolute error ranging from 1.2 °C (autumn and summer) to 1.3 °C (winter and spring), and corresponding root-mean-square error in the range 1.6 °C to 1.7 °C. The R2 and concordance correlation coefficient (Pc ) values were also above 0.8 in each season indicating predictions were strongly correlated to the validation data. On an hourly basis, errors tended to be highest during the late afternoons in spring and summer from 3 pm to 6 pm, particularly for the coastal areas of Western Australia. The mapping system was trialled over a 21-day period from 1 June 2020 to 21 June 2020 with majority of maps completed within 28-min of AWS site observations being recorded. All outputs were displayed in a web mapping application to exemplify a real-time application of the outputs. This study found that the methods employed would be highly suited for similar applications requiring real-time processing and delivery of climate data at high spatiotemporal resolutions across a considerably large land mass.

Development of a conceptual GIS model to support management decision making

Making (identifying) the right decision, just making a decisionis a choice from a variety of acceptable alternatives, depending on various factors, one in which the effective value is optimized. If a decision maker (DM) or a decision support system (DSS) can isolate a control parameter with unconditional preference that is relevant and most fully characterizing the properties of the system (process, object), it can be considered an objective function, subject to certain restrictive (by an acceptable set, resources) conditions. A similar problem is single-criterion, solved by classical methods of the decision theory. Decision-making tasks can be solved by a sequence of steps, a procedure: generation of acceptable options, selection of relevant criteria, search for an acceptable choice, its evaluation. In solving the problems of the decision support system (DSS), the decision-maker is usually involved - an expert, an analyst with competencies in the subject area and evaluating options, a consultant who helps to formulate acceptable options. The article discusses the development of a conceptual model of a geographic information system to support managerial decision-making. The basis of such GIS is an automated mapping system.

Automated Mapping of Convective Clouds (AMCC) Thermodynamical, Microphysical, and CCN Properties from SNPP/VIIRS Satellite Data

AbstractThe advent of the Visible Infrared Imager Radiometer Suite (VIIRS) on board the Suomi NPP (SNPP) satellite made it possible to retrieve a new class of convective cloud properties and the aerosols that they ingest. An automated mapping system of retrieval of some properties of convective cloud fields over large areas at the scale of satellite coverage was developed and is presented here. The system is named Automated Mapping of Convective Clouds (AMCC). The input is level-1 VIIRS data and meteorological gridded data. AMCC identifies the cloudy pixels of convective elements; retrieves for each pixel its temperature T and cloud drop effective radius re; calculates cloud-base temperature Tb based on the warmest cloudy pixels; calculates cloud-base height Hb and pressure Pb based on Tb and meteorological data; calculates cloud-base updraft Wb based on Hb; calculates cloud-base adiabatic cloud drop concentrations Nd,a based on the T–re relationship, Tb, and Pb; calculates cloud-base maximum vapor supersaturation S based on Nd,a and Wb; and defines Nd,a/1.3 as the cloud condensation nuclei (CCN) concentration NCCN at that S. The results are gridded 36 km × 36 km data points at nadir, which are sufficiently large to capture the properties of a field of convective clouds and also sufficiently small to capture aerosol and dynamic perturbations at this scale, such as urban and land-use features. The results of AMCC are instrumental in observing spatial covariability in clouds and CCN properties and for obtaining insights from such observations for natural and man-made causes. AMCC-generated maps are also useful for applications from numerical weather forecasting to climate models.

Use of Novel Electrogram "Lumipoint" Algorithm to Detect Critical Isthmus and Abnormal Potentials for Ablation in Ventricular Tachycardia.

This study reports the use of a novel "Lumipoint" algorithm in ventricular tachycardia (VT) ablation. Automatic mapping systems aid rapid acquisition of activation maps. However, they may annotate farfield rather than nearfield signal in low voltage areas, making maps difficult to interpret. The Lumipoint algorithm analyzes the complete electrogram tracing and therefore includes nearfield signals in its analysis. Twenty-two patients with ischemic cardiomyopathy and 5 with dilated cardiomyopathy underwent mapping using the ultra-high density Rhythmia system. Lumipoint algorithms were applied retrospectively. In all left ventricular substrate maps, changing the window of interest to the post-QRS phase automatically identified late potentials. In 25 of 27 left ventricular VT activation maps, a minimum spatial window of interest correctly identified the VT isthmus as seen by the manually annotated map, entrainment, and response to ablation. In 6 maps, the algorithm identified the isthmus where the standard automatically annotated map did not. The Lumipoint algorithm automatically highlights areas with electrograms having specific characteristics or timings. This can identify late and fractionated potentials and regions that exhibit discontinuous activation, as well as the isthmus of a VT circuit. These features may enhance human interpretation of the electrogram signals during a case, particularly where the circuit lies in partial scar with low amplitude nearfield signals and potentially allow a more targeted ablation strategy.

The use of a high-resolution mapping system may facilitate standard clinical practice in VE and VT ablation.

First experiences using a 64-electrode mini-basket catheter (BC) paired with an automatic mapping system (Rhythmia™) for catheter ablation (CA) of ventricular ectopy (VE) and ventricular tachycardia (VT) have been reported. We aimed to evaluate (1) differences in ventricular access for the BC and (2) benefit of this technology in the setting of standard clinical practice. Patients (pts) undergoing CA for VE or VT using the Intellamap Orion™ paired with the Rhythmia™ automated-mapping system were included in this study. For LV access, transseptal and retrograde access were compared. All 32 pts (29 men, age 63 ± 15years) underwent CA for VE (17 pts) or VT (15 pts). For mapping of VE originating from the left ventricle (LV) in 10 out of 13 pts, a transaortic access was feasible. The predominant access for CA of VT was transaortic (5/7). Feasibility and safety seem to be equal. The total procedure time was 179.1 ± 21.2min for VE ablation and 212.0 ± 71.7min for VT ablation (p = 0.177). For VE, an acquisition of 1602 ± 1672 map points and annotation of 140 ± 98 automated mapping points sufficed to abolish VE in all pts. During a 6-month follow-up (FU) after CA for VE, a VE burden reduction from 18.5 ± 2.1% to 2.8 ± 2.2% (p = 0.019) was achieved. In VT pts, one patient showed recurrence of sustained VT episodes during FU. Use of a high-resolution mapping system for VE/VT CA potentially facilitates revelation of VE origin and VT circuits in the setting of standard clinical practice. Feasibility and safety of a venous, transaortic, transseptal, or a combined approach seem to be equal.

Automatic wireless mapping and tracking system for indoor location

Automatic vehicle tracking systems ease the completion of numerous tasks in different fields. Moreover they can automatically capture information, this feature allows to perform location tasks. These systems can be implemented at airports, in shopping centers and in other large buildings; in this way, wireless network scans will serve as a basis for the creation of signal maps that can be used in indoor location systems. This work proposes an automatic people tracking system which also allows to map Wi-Fi networks in order to localize people indoor. In order to operate the system, information on vehicle movement was used to capture signal maps, with the aim of reducing the need to perform manual calibration and thus, improving the updating of information. The final location is determined by combining information provided by wireless networks, Bayesian networks are employed for this task.

Characteristics of Single-Loop Macroreentrant Biatrial Tachycardia Diagnosed by Ultrahigh-Resolution Mapping System.

Biatrial tachycardia (BiAT) is a rare form of atrial macroreentrant tachycardia, in which both atria form a critical part of the circuit. We aimed to identify the characteristics and precise circuits of single-loop macroreentrant BiATs. We identified 8 patients (median age, 59.5 years old) with 9 BiATs in a cohort of 336 consecutive patients from 2 institutions who had undergone AT catheter ablation using an automatic ultrahigh-resolution mapping system. Seven of the 8 patients had a history of persistent AF ablation, including septal or anterior left atrium ablation before developing BiAT. One of the 8 patients had a history of an atrial septal patch closure with a massively enlarged right atrium. Nine ATs (median cycle length, 334 ms; median 12 561 points in the left atrium; 8814 points in the right atrium) were diagnosed as single-loop macroreentrant BiATs. We observed 3 types of BiAT (1) BiAT with a perimitral and peritricuspid reentrant circuit (n=3), (2) BiAT using the right atrium septum and a perimitral circuit (n=3), and (3) BiAT using only the left atrium and right atrium septum (n=3). Catheter ablation successfully terminated 8 of the 9 BiATs. All patients who developed BiAT had an electric obstacle on the anteroseptal left atrium, primarily from prior ablation lesions. In this situation, mapping of both atria should be considered during AT. Because 3 types of single-loop BiAT were observed, ablation strategies should be adjusted to the type of BiAT circuit.

Improving the visibility of library resources via mapping library subject headings to Wikipedia articles

PurposeLinking libraries and Wikipedia can significantly improve the quality of services provided by these two major silos of knowledge. Such linkage would enrich the quality of Wikipedia articles and at the same time increase the visibility of library resources. To this end, the purpose of this paper is to describe the design and development of a software system for automatic mapping of FAST subject headings, used to index library materials, to their corresponding articles in Wikipedia.Design/methodology/approachThe proposed system works by first detecting all the candidate Wikipedia concepts (articles) occurring in the titles of the books and other library materials which are indexed with a given FAST subject heading. This is then followed by training and deploying a machine learning (ML) algorithm designed to automatically identify those concepts that correspond to the FAST heading. In specific, the ML algorithm used is a binary classifier which classifies the candidate concepts into either “corresponding” or “non-corresponding” categories. The classifier is trained to learn the characteristics of those candidates which have the highest probability of belonging to the “corresponding” category based on a set of 14 positional, statistical, and semantic features.FindingsThe authors have assessed the performance of the developed system using standard information retrieval measures of precision, recall, and F-score on a data set containing 170 FAST subject headings manually mapped to their corresponding Wikipedia articles. The evaluation results show that the developed system is capable of achieving F-scores as high as 0.65 and 0.99 in the corresponding and non-corresponding categories, respectively.Research limitations/implicationsThe size of the data set used to evaluate the performance of the system is rather small. However, the authors believe that the developed data set is large enough to demonstrate the feasibility and scalability of the proposed approach.Practical implicationsThe sheer size of English Wikipedia makes the manual mapping of Wikipedia articles to library subject headings a very labor-intensive and time-consuming task. Therefore, the aim is to reduce the cost of such mapping and integration.Social implicationsThe proposed mapping paves the way for connecting libraries and Wikipedia as two major silos of knowledge, and enables the bi-directional movement of users between the two.Originality/valueTo the best of the authors’ knowledge, the current work is the first attempt at automatic mapping of Wikipedia to a library-controlled vocabulary.

Rapid Identification of Cortical Motor Areas in Rodents by High-Frequency Automatic Cortical Stimulation and Novel Motor Threshold Algorithm.

Cortical stimulation mapping is a valuable tool to test the functional organization of the motor cortex in both basic neurophysiology (e.g., elucidating the process of motor plasticity) and clinical practice (e.g., before resecting brain tumors involving the motor cortex). However, compilation of motor maps based on the motor threshold (MT) requires a large number of cortical stimulations and is therefore time consuming. Shortening the time for mapping may reduce stress on the subjects and unveil short-term plasticity mechanisms. In this study, we aimed to establish a cortical stimulation mapping procedure in which the time needed to identify a motor area is reduced to the order of minutes without compromising reliability. We developed an automatic motor mapping system that applies epidural cortical surface stimulations (CSSs) through one-by-one of 32 micro-electrocorticographic electrodes while examining the muscles represented in a cortical region. The next stimulus intensity was selected according to previously evoked electromyographic responses in a closed-loop fashion. CSS was repeated at 4 Hz and electromyographic responses were submitted to a newly proposed algorithm estimating the MT with smaller number of stimuli with respect to traditional approaches. The results showed that in all tested rats (n = 12) the motor area maps identified by our novel mapping procedure (novel MT algorithm and 4-Hz CSS) significantly correlated with the maps achieved by the conventional MT algorithm with 1-Hz CSS. The reliability of the both mapping methods was very high (intraclass correlation coefficients ≧0.8), while the time needed for the mapping was one-twelfth shorter with the novel method. Furthermore, the motor maps assessed by intracortical microstimulation and the novel CSS mapping procedure in two rats were compared and were also significantly correlated. Our novel mapping procedure that determined a cortical motor area within a few minutes could help to study the functional significance of short-term plasticity in motor learning and recovery from brain injuries. Besides this advantage, particularly in the case of human patients or experimental animals that are less trained to remain at rest, shorter mapping time is physically and mentally less demanding and might allow the evaluation of motor maps in awake individuals as well.

Challenges and Pitfalls of Entrainment Mapping of Ventricular Tachycardia: Ten Illustrative Concepts.

The use and interpretation of entrainment mapping, or continuous resetting, of a reentrant tachycardia has been regarded as the gold standard for delineation of the components of a reentrant circuit.1,2 The response during and after overdrive pacing, whereby 2 wavefronts enter the circuit antidromically (with fusion) and orthodromically, is used to confirm reentry as the arrhythmia mechanism and determine the relationship of the pacing site to the circuit. The fulfillment of classical criteria outlined by Waldo and colleagues3,4 were synthesized into an anatomic concept for scar-related ventricular tachycardia (VT) by Stevenson et al5 to portray the structural and architectural basis of circuit conduction meandering between regions of fibrosis. In this construct, a central corridor, or protected isthmus, is bordered between 2 regions of dense scar with a single entrance that is distinct from a single exit, which yields the QRS morphology.6 This reentrant model has been central to our current mechanistic understanding of scar-mediated VT and is critically important for differentiating critical sites from bystander sites and regions that are unlikely to interrupt or eliminate reentry.7 However, the nature of reentrant VT in man is more complex than our idealized working construct for many reasons. In clinical practice, the majority of VT is hemodynamically unstable, which precludes the ability to perform entrainment mapping and activation mapping of the entire circuit.8 Differences in the circuit between patients with untolerated and tolerated VT are not well understood. VT circuits are 3 dimensionally complex with transmural conduction and circuit conduction is unlikely to be planar, as depicted by electroanatomic mapping of the myocardial surface. Exits may be multiple9,10 and patterns other than loop reentry around scar are likely. Channels of preserved myocardium are frequently not “normal” in voltage (>1.5 mV) …

Evaluation of a novel high-resolution mapping system for catheter ablation of ventricular arrhythmias

The mapping of ventricular arrhythmias in humans using a minibasket 64-electrode catheter paired with a novel automatic mapping system (Rhythmia) has not been evaluated. The purpose of this study was to evaluate the safety and efficacy of mapping ventricular arrhythmias and clinical outcomes after ablation using this system. Electroanatomic maps for ventricular arrhythmias were obtained during 20 consecutive procedures in 19 patients (12 with ventricular tachycardia [VT] and 2 with ventricular ectopy [VE]). High-density maps were acquired using automatic beat acceptance and automatic system annotation of electrograms. Forty-seven electroanatomic maps (including 3 right ventricular and 9 epicardial maps) were obtained. Left ventricular endocardial mapping by transseptal (n = 13) and/or transaortic (n = 11) access was safe with no complications related to the minibasket catheter. VT substrate maps (n = 14; median 10,184 points) consistently demonstrated late potentials with high resolution. VT activation maps (n = 25; median 6401 points) obtained by automatic annotation included 7 complete maps (covering ≥90% of the tachycardia cycle length) in 5 patients in whom the entire VT circuit was accurately visualized. VE timing maps (n = 8) successfully localized the origin of VEs in all, with all accepted beats consistent with clinical VEs. Over a median follow-up of 10 months, no arrhythmia recurrence was noted in 75% after VT ablation and 86% after VE ablation. In this first human experience for ventricular arrhythmias using this system, ultra-high-density maps were created rapidly and safely, with a reliable automatic annotation of VT and consistent recording of abnormal electrograms. Medium-term outcomes after ablation were encouraging. Further larger studies are needed to validate these findings.

AN AUTOMATIC UAV MAPPING SYSTEM FOR SUPPORTING UN (UNITED NATIONS) FIELD OPERATIONS

Abstract. The United Nations (UN) has performed field operations worldwide such as peacekeeping or rescue missions. When such an operation is needed, the UN dispatches an operation team usually with a GIS (Geographic Information System) customized to a specific operation. The base maps for the GIS are generated mostly with satellite images which may not retain a high resolution and the current situation. To build an up-to-date high resolution map, we propose a UAV (unmanned aerial vehicle) based automatic mapping system, which can operate in a fully automatic way from the data acquisition of sensory data to the data processing for the generation of the geospatial products such as a mosaicked orthoimage of a target area. In this study, we analyse the requirements for UN field operations, suggest a UAV mapping system with an operation scenario, and investigate the applicability of the system. With the proposed system, we can construct a tailored GIS with up-to-date and high resolution base maps for a specific operation efficiently.

AN AUTOMATIC UAV MAPPING SYSTEM FOR SUPPORTING UN (UNITED NATIONS) FIELD OPERATIONS

The United Nations (UN) has performed field operations worldwide such as peacekeeping or rescue missions. When such an operation is needed, the UN dispatches an operation team usually with a GIS (Geographic Information System) customized to a specific operation. The base maps for the GIS are generated mostly with satellite images which may not retain a high resolution and the current situation. To build an up-to-date high resolution map, we propose a UAV (unmanned aerial vehicle) based automatic mapping system, which can operate in a fully automatic way from the data acquisition of sensory data to the data processing for the generation of the geospatial products such as a mosaicked orthoimage of a target area. In this study, we analyse the requirements for UN field operations, suggest a UAV mapping system with an operation scenario, and investigate the applicability of the system. With the proposed system, we can construct a tailored GIS with up-to-date and high resolution base maps for a specific operation efficiently.

Utility of a Novel Rapid High-Resolution Mapping System in the Catheter Ablation of Arrhythmias: An Initial Human Experience of Mapping the Atria and the Left Ventricle

ObjectivesThis study sought to assess the clinical efficacy, safety, and clinical utility of a novel electroanatomical mapping system. BackgroundA new mapping system capable of rapidly acquiring detailed maps based on automatic annotation of thousands of points was recently released for clinical use. This is the first description of its utility in humans. MethodsThe first consecutive 20 cases (7 atrial tachycardia, 8 atrial fibrillation, 3 ventricular tachycardia, and 2 ventricular ectopic beat ablations) were analyzed. The system uses a bidirectional deflectable basket catheter with 64 closely spaced mini-electrodes. It automatically accepts and annotates electrograms when a number of predefined criteria are met. ResultsThirty right atrial maps were acquired in 11 (4 to 15) min, consisting of 7,220 (3,467 to 10,947) points, 22 left atrial maps in 11 (6 to 19) min, consisting of 7,818 (4,379 to 12,262) points and 10 left ventricular maps in 37 (14 to 43) min, consisting of 8,709 (2,605 to 15,514) points. The mini-basket catheter could reach all areas of interest without deflectable sheaths. No embolic events, bleeding complications, or endocardial structure damage were observed. Correction of the automatic annotation was performed in 0.02% of points in 4 of 62 maps. The system revealed re-entry circuits of atrial tachyarrhythmias, identified gaps on linear lesions, and identified and correctly annotated the clinical ventricular ectopic beats and channels of slow conduction within ventricular scar. ConclusionsThe novel automatic mapping system was rapid, safe, and efficacious in mapping a variety of cardiac arrhythmias in humans. Further clinical research is needed to optimize its use in the ablation of complex arrhythmias.

An Extensible System for Multilevel Automatic Data Partition and Mapping

Automatic data distribution is a key feature to obtain efficient implementations from abstract and portable parallel codes. We present a highly efficient and extensible runtime library that integrates techniques for automatic data partition and mapping. It uses a novel approach to define an abstract interface and a plug-in system to encapsulate different types of regular and irregular techniques, helping to generate codes which are independent of the exact mapping functions selected. Currently, it supports hierarchical tiling of arrays with dense and stride domains, that allows the implementation of both data and task parallelism using a SPMD model. It automatically computes appropriate domain partitions for a selected virtual topology, mapping them to available processors with static or dynamic load-balancing techniques. Our library also allows the construction of reusable communication patterns that efficiently exploit MPI communication capabilities. The use of our library greatly reduces the complexity of data distribution and communication, hiding the details of the underlying architecture. The library can be used as an abstract layer for building generic tiling operations as well. Our experimental results show that the use of this library allows to achieve similar performance as carefully-implemented manual versions for several, well-known parallel kernels and benchmarks in distributed and multicore systems, and substantially reduces programming effort.

A system for large-scale automatic traffic sign recognition and mapping

Abstract. We present a system for the large-scale automatic traffic signs recognition and mapping and experimentally justify design choices made for different components of the system. Our system works with more than 140 different classes of traffic signs and does not require labor-intensive labelling of a large amount of training data due to the training on synthetically generated images. We evaluated our system on the large dataset of Russian traffic signs and made this dataset publically available to encourage future comparison.