Hilario Gómez-Moreno

Road-Sign Detection and Recognition Based on Support Vector Machines

This paper presents an automatic road-sign detection and recognition system based on support vector machines (SVMs). In automatic traffic-sign maintenance and in a visual driver-assistance system, road-sign detection and recognition are two of the most important functions. Our system is able to detect and recognize circular, rectangular, triangular, and octagonal signs and, hence, covers all existing Spanish traffic-sign shapes. Road signs provide drivers important information and help them to drive more safely and more easily by guiding and warning them and thus regulating their actions. The proposed recognition system is based on the generalization properties of SVMs. The system consists of three stages: 1) segmentation according to the color of the pixel; 2) traffic-sign detection by shape classification using linear SVMs; and 3) content recognition based on Gaussian-kernel SVMs. Because of the used segmentation stage by red, blue, yellow, white, or combinations of these colors, all traffic signs can be detected, and some of them can be detected by several colors. Results show a high success rate and a very low amount of false positives in the final recognition stage. From these results, we can conclude that the proposed algorithm is invariant to translation, rotation, scale, and, in many situations, even to partial occlusions

Goal Evaluation of Segmentation Algorithms for Traffic Sign Recognition

This paper presents a quantitative comparison of several segmentation methods (including new ones) that have successfully been used in traffic sign recognition. The methods presented can be classified into color-space thresholding, edge detection, and chromatic/achromatic decomposition. Our support vector machine (SVM) segmentation method and speed enhancement using a lookup table (LUT) have also been tested. The best algorithm will be the one that yields the best global results throughout the whole recognition process, which comprises three stages: 1) segmentation; 2) detection; and 3) recognition. Thus, an evaluation method, which consists of applying the entire recognition system to a set of images with at least one traffic sign, is attempted while changing the segmentation method used. This way, it is possible to observe modifications in performance due to the kind of segmentation used. The results lead us to conclude that the best methods are those that are normalized with respect to illumination, such as RGB or Ohta Normalized, and there is no improvement in the use of Hue Saturation Intensity (HSI)-like spaces. In addition, an LUT with a reduction in the less-significant bits, such as that proposed here, improves speed while maintaining quality. SVMs used in color segmentation give good results, but some improvements are needed when applied to achromatic colors.

Traffic sign recognition system for inventory purposes

This paper describes the evaluation of the characteristics of a real automatic traffic sign detection system. The objective of this review is to provide the basis of quality of a whole system, which is capable of identifying the different signs that can be found in route. At the moment, our work is concerned with the developing of an inventory system capable to get a complete catalog of all the traffic signs and their corresponding state information. The paper analyzes exhaustively the different problems that can appear in real environments and shows how the system implemented overcomes all these difficulties with a high success. The flexibility of the system allows it to run new algorithms even though several of them can be run in parallel and, on the other hand, it is relatively easy to change the training traffic sign according to the circumstances: urban or non-urban environments and traffic signs from different countries.

Shape classification algorithm using support vector machines for traffic sign recognition

In this paper, a new algorithm for traffic sign recognition is presented. It is based on a shape detection algorithm that classifies the shape of the content of a sign using the capabilities of a Support Vector Machine (SVM). Basically, the algorithm extracts the shape inside a traffic sign, computes the projection of this shape and classifies it into one of the shapes previously trained with the SVM. The most important advances of the algorithm is its robustness against image rotation and scaling due to camera projections, and its good performance over images with different levels of illumination. This work is part of a traffic sign detection and recognition system, and in this paper we will focus solely on the recognition step.

Traffic sign shape classification based on Support Vector Machines and the FFT of the signature of blobs

In many traffic sign recognition systems, one of the main tasks is the classification of the shape of the blob, which is intended to simplify the recognition process. In this paper, we have developed a new shape classification algorithm based on Support Vector Machines classifiers and the FFT of the signature of the blob. The FFT of the signature yields invariance to object scalings and rotations. Furthermore, the FFT is the vector input to the classifier. This classifier is trained to cope with projection deformations and occlusions. The algorithm has been tested under adverse conditions, such as geometric distortions, i.e. scaling, rotations and projection deformations, and occlusions. The experimental results show good robustness when the system is working with real, outdoor road images.

Edge Detection in Noisy Images Using the Support Vector Machines

In this paper, a new method for edge detection in presence of impulsive noise based into the use of Support Vector Machines (SVM) is presented. This method shows how the SVM can detect edge in an efficient way. The noisy images are processed in two ways, first reducing the noise by using the SVM regression and then performing the classification using the SVM classification. The results presented show that this method is better than the classical ones when the images are affected by impulsive noise and, besides, it is well suited when the images are not noisy.

Background pixel classification for motion detection in video image sequences

Simoncelli & Heeger studied how the motion is processed in humans (V1 and MT areas) and proposed a model based on neural populations that extract the local motion structure through local competition of MT like cells. In this paper we present a neural structure that works as dynamic filter on the top of this MT layer and can take advantage of the neural population coding that is supposed to be present in the MT cortical processing areas. The test bed application addressed in this work is an automatic watch up system for the rear-view mirror blind spot. The segmentation of overtaking cars in this scenario can take full advantage of the motion structure of the visual field provided that the ego-motion of the host car induces a global motion pattern whereas an overtaking car produces a motion pattern highly contrasted with this global ego-motion field.

Extracting illumination from images by using the wavelet transform

We present an efficient way to extract the illumination from the images by exploring only the low frequencies into them jointly with the use of the illumination model from the homomorphic filter. This illumination information may be used to code images in a sequence with only illumination changes. We take an image from the sequence as a reference. Any other image to be coded is approximated by changing the low pass band of the discrete wavelet transform of the reference with that from the image to be coded. To improve the recovering process it is possible to modify the high frequencies into the reference image according to the differences between the illumination from the reference and from the image to be recovered. The results obtained show us a new way to code illumination changes when used with video sequences.

Continuous variance estimation in video surveillance sequences with high illumination changes

Continuous estimation of signal statistics is an important issue in many video processing systems, such as motion detection in surveillance applications. In this paper we demonstrate how results of classical expressions for variance estimation decrease in accuracy when dealing with sequences containing high illumination variations. The paper also proposes a new estimation method, and shows how, under such conditions, the accuracy of the proposed method produces better results whilst maintaining performance in scenarios with smaller changes, thus improving the motion detection stage of a video surveillance system.

A “Salt and Pepper” Noise Reduction Scheme for Digital Images Based on Support Vector Machines Classification and Regression

We present a new impulse noise removal technique based on Support Vector Machines (SVM). Both classification and regression were used to reduce the “salt and pepper” noise found in digital images. Classification enables identification of noisy pixels, while regression provides a means to determine reconstruction values. The training vectors necessary for the SVM were generated synthetically in order to maintain control over quality and complexity. A modified median filter based on a previous noise detection stage and a regression-based filter are presented and compared to other well-known state-of-the-art noise reduction algorithms. The results show that the filters proposed achieved good results, outperforming other state-of-the-art algorithms for low and medium noise ratios, and were comparable for very highly corrupted images.