No abstract

In this paper we discuss theoretical foundations and a practical realization of a real-time traffic sign detection, tracking and recognition system operating on board of a vehicle. In the proposed framework a generic detector refinement procedure based on a mean shift clustering is introduced. This technique is shown to improve the detection accuracy and reduce the number of false positives for a broad class of object detectors for which a soft response's confidence can be sensibly measured. Track of an already established candidate is maintained over time using an instance-specific tracking function that encodes the relationship between a unique feature representation of the target object and the affine distortions it is subject to. We show that this function can be learned on-the-fly via regression from random transformations applied to the image of the object in known pose. Secondly, we demonstrate its capability of reconstructing the full-face view of a sign from substantial viewangles. In the classification stage a concept of a similarity measure learned from image pairs is discussed and its realization using SimBoost, a novel version of AdaBoost algorithm, is analyzed. Suitability of the proposed method for solving multi-class traffic sign classification problems is shown experimentally for different image representations. Overall performance of the entire system is evaluated based on a prototype C++ implementation. Illustrative output generated by this demo application is provided as a supplementary material attached to this paper. Journal of Machine Vision & ApplicationsThis work may not be copied or reproduced in whole or in part for any commercial purpose. Permission to copy in whole or in part without payment of fee is granted for nonprofit educational and research purposes provided that all such whole or partial copies include the following: a notice that such copying is by permission of Mitsubishi Electric Research Laboratories, Inc.; an acknowledgment of the authors and individual contributions to the work; and all applicable portions of the copyright notice. Copying, reproduction, or republishing for any other purpose shall require a license with payment of fee to Mitsubishi Electric Research Laboratories, Inc. All rights reserved. In-Vehicle Camera Traffic Sign Detection and RecognitionAndrzej Ruta · Fatih Porikli · Yongmin Li · Shintaro WatanabeReceived: dd-mm-yyyy / Accepted: dd-mm-yyyyAbstract In this paper we discuss theoretical foundations and a practical realization of a real-time traffic sign detection, tracking and recognition system operating on board of a vehicle. In the proposed framework a generic detector refinement procedure based on a mean shift clustering is introduced. This technique is shown to improve the detection accuracy and reduce the number of false positives for a broad class of object detectors for which a soft response's confidence can be sensibly measured. Track of an already established candidate is maintained over time using an instance-specific tracking fun...

Abstract-In this paper a comprehensive approach to the recognition of traffic signs from video input is proposed. A trained attentive classifier cascade is used to scan the scene in order to quickly establish regions of interest (ROI). Sign candidates within ROIs are captured by detecting the instances of equiangular polygons using a Hough Transform-style shape detector. To ensure a stable tracking of the likely traffic signs, especially in cluttered background, we propose a Pixel Relevance Model, where the pixel relevance is defined as a confidence measure for a pixel being part of a sign's contour. The relevance of the hypothesized contour pixels is updated dynamically within a small search region maintained by a Kalman Filter, which ensures faster computation. Gradient magnitude is used as an observable evidence for this update process. In the classification stage, a temporally integrated template matching technique based on the class-specific discriminative local region representation of an image is adopted. We have evaluated the proposed approach on a large database of 135 traffic signs and numerous real traffic video sequences. A recognition accuracy of over 93% in near real-time has been achieved.

Real-time road sign recognition has been of great interest for many years. This problem is often addressed in a two-stage procedure involving detection and classification. In this paper a novel approach to sign representation and classification is proposed. In many previous studies focus was put on deriving a set of discriminative features from a large amount of training data using global feature selection techniques e.g. Principal Component Analysis or AdaBoost. In our method we have chosen a simple yet robust image representation built on top of the Colour Distance Transform (CDT). Based on this representation, we introduce a feature selection algorithm which captures a variable-size set of local image regions ensuring maximum dissimilarity between each individual sign and all other signs. Experiments have shown that the discriminative local features extracted from the template sign images enable minimum-distance classification with error rate not exceeding 7%.

No abstract

In this paper we focus on realistic clustering problems where the input data is high-dimensional and the clusters have complex, multimodal distribution. In this challenging setting the conventional methods, such as k-centers family, hierarchical clustering or those based on model fitting, are inefficient and typically converge far from the globally optimal solution. As an alternative, we propose a novel unsupervised learning approach which is based on the compressive sensing paradigm. The key idea underlying our algorithm is to monitor the distance between the test sample and its principal projection in each cluster, and continue re-assigning it to the cluster yielding the smallest residual. As a result, we obtain an iterative procedure which, under the compressive assumptions, minimizes the total reconstruction error of all samples from their nearest clusters. To evaluate the proposed approach, we have conducted a series of experiments involving various image collections where the task was to automatically group similar objects. Comparison of the obtained results with those yielded by the state-of-the-art clustering methods provides evidence for high discriminative power of our algorithm. International Conference on Machine Learning and Applications (ICMLA)This work may not be copied or reproduced in whole or in part for any commercial purpose. Permission to copy in whole or in part without payment of fee is granted for nonprofit educational and research purposes provided that all such whole or partial copies include the following: a notice that such copying is by permission of Mitsubishi Electric Research Laboratories, Inc.; an acknowledgment of the authors and individual contributions to the work; and all applicable portions of the copyright notice. Copying, reproduction, or republishing for any other purpose shall require a license with payment of fee to Mitsubishi Electric Research Laboratories, Inc. All rights reserved. Abstract-In this paper we focus on realistic clustering problems where the input data is high-dimensional and the clusters have complex, multimodal distribution. In this challenging setting the conventional methods, such as k-centers family, hierarchical clustering or those based on model fitting, are inefficient and typically converge far from the globally optimal solution. As an alternative, we propose a novel unsupervised learning approach which is based on the compressive sensing paradigm. The key idea underlying our algorithm is to monitor the distance between the test sample and its principal projection in each cluster, and continue re-assigning it to the cluster yielding the smallest residual. As a result, we obtain an iterative procedure which, under the compressive assumptions, minimizes the total reconstruction error of all samples from their nearest clusters. To evaluate the proposed approach, we have conducted a series of experiments involving various image collections where the task was to automatically group similar objects. Comparison of the obtained results with thos...

No abstract

Development of FPGA-based, network-enabled embedded systems in Register Transfer Level hardware description languages is tedious. Despite the automation of this process with numerous EDA tools available, no well-established design patterns exist. Moreover, the entire production cycle requires appropriate theoretical background and hardware design intuition from the developer which discourages the software community. To improve productivity and minimize time to market when assembling such systems, we propose a new hardware/software co-design approach to building reconfigurable hardware web services. The proposed integrated development platform features a programmable FPGA board where computations of different nature and purpose are logically distributed among a sequential soft-core processor program, a massively parallel accelerator and an independent communication module that handles remote clients' requests. Our second contribution is a set of tools that make the development of the aforementioned services essentially a software design undertaking with the extensive use of high-level programming languages. The platform has been tuned to act as a flexible runtime environment for image processing services, thus providing functionality of an intelligent camera. Two example services built from scratch according to the new methodology are discussed. Reduced development time and significant performance gain observed prove validity of the proposed approach and unveil a large potential of the assembled prototype.