Chokri Souani

Efficient algorithm for automatic road sign recognition and its hardware implementation

The automatic detection of road signs is an application that alerts the vehicle’s driver of the presence of signals and invites him to react on time in the aim to avoid potential traffic accidents. This application can thus improve the road safety of persons and vehicles traveling in the road. Several techniques and algorithms allowing automatic detection of road signs are developed and implemented in software and do not allow embedded application. We propose in this work an efficient algorithm and its hardware implementation in an embedded system running in real time. In this paper we propose to implement the application of automatic recognition of road signs in real time by optimizing the techniques used in different phases of the recognition process. The system is implemented in a Virtex4 FPGA family which is connected to a camera mounted in the moving vehicle. The system can be integrated into the dashboard of the vehicle. The performance of the system shows a good compromise between speed and efficiency.

Embedded system for road sign detection using MicroBlaze

In this paper, we present implementation of road sign detection application in an embedded real time system. The target hardware is a Xilinx MicroBlaze soft-processor. The algorithm is described using C language and compiled with the SDK and EDK tools targeting the VirtexML605 FPGA. The input video is a real scene acquired by a digital camera with resolution of 480×640 pixels. Adequate architecture generated of the MicroBlaze with adequate communication with the system memory, allows real time execution of the application. Experimental results on synthetic and real data show that our implementation successfully runs in real time with a low occupation of resources, while maintaining comparable functionality to version existing solutions based on FPGA. The proposed hardware platform is faster compared to existing solutions based DSP.

Novel approach for real time detection and classification based on template matching in video

In this paper, we present a new approach for the detection and classification of real time road signs from video. The proposed system is composed of two processing stages: detection and classification signs. The system has detected road signs by color and shape feature of segmentation color in HSV color space, especially red and blue. The detected signs are then classified by method template matching, such as circular, squared and triangular shapes. The road signs are classified in one of the following categories by Combining color and information about shape: danger, obligation, prohibition or information. As proposed system input is a video with a resolution of 1360×800 pixels. For detection phase road signs, has a high detection performance up to 92 % and for classification is rated 96% in our experiments. The proposed system also proves to be reliable and suitable for real-time processing.

Real-time embedded system for traffic sign recognition based on ZedBoard

AbstractThis paper presents a design methodology of a real-time embedded system that processes the detection and recognition of road signs while the vehicle is moving. An efficient algorithm was proposed, which operates in two processing steps: the detection and the recognition. Regions of interest were extracted by using the Maximally Stable Extremal Regions Method. For the recognition phase, Oriented FAST and Rotated BRIEF features were used. A hardware system based on the Xilinx Zynq platform was developed. The designed system can achieve real-time video processing while assuring constraints and a high-level accuracy in terms of detection and recognition rates.

Effect of color spaces on video segmentation performances

Color segmentation is a preliminary step in many application computer vision systems today, as the detection of human movement, recognition of road traffic signs and video intelligent. Detection stage performance is therefore closely linked to the results obtained from the color segmentation. Detection and recognition automatic road traffic signs are applied in the color spaces RGB, HSV, and HSI. We present in this paper a comparative study on the color detection of road signs by thresholding according to the color spaces: RGB, HSV and HSI We proceed after that to verify the results of the detection rate and false detection rate obtained for each color space. The results have been validated on video under different lighting conditions.

Real-time recognition of road traffic signs in video scenes

This paper presents a new algorithm for the detection and classification of real-time road traffic signs in the video. Our system is able to detect and classify triangular, circular, and octagonal signs of red and blue colors. The proposed system operates into two processing steps: (1) detection and (2) classification road signs. The system has detected candidate regions as Maximally Stable Extremal Regions (MSERs) in HSV color space with available robustness to variations in lighting conditions, especially red and blue. The detected candidate regions were then classified as method template matching, such as circular, octagonal, and triangular shapes. The proposed system is operating under a range of weather conditions and recognizes all classes of road traffic sign database. Results show a high success rate. In fact, the system maintains high performance for detection and classification steps whose F-measure are set to 0.95 and 0.92 respectively. We conclude, from these results, that the proposed system is invariant to rotation, translation, scale, even to partial occlusions. Moreover, results prove that the system is suitable and reliable for real-time processing.

DSP implementation and performances evaluation of 1D and 2D DWT using the lifting scheme

This paper presents the implementation of a lifting scheme-based DWT architecture on a digital signal processor (DSP) (TMS320C6713) core and the simulation using MATLAB. The 5/3 and 9/7 wavelet filters used in JPEG 2000 are both implanted and executed by the DSP processor for comparisons. The algorithms proposed are optimized at source code level and memory usage. The execution time for performing both DWTs is measured for 1D and 2D-DWT for different number of level, depending on on-chip and off-chip memory.

Novel Technique for 3D Face Recognition Using Anthropometric Methodology

This manuscript presents an improved system research that can detect and recognize the person in 3D space automatically and without the interaction of the peoples faces. This system is based not only on a quantum computation and measurements to extract the vector features in the phase of characterization but also on learning algorithm using SVM to classify and recognize the person. This research presents an improved technique for automatic 3D face recognition using anthropometric proportions and measurement to detect and extract the area of interest which is unaffected by facial expression. This approach is able to treat incomplete and noisy images and reject the non-facial areas automatically. Moreover, it can deal with the presence of holes in the meshed and textured 3D image. It is also stable against small translation and rotation of the face. All the experimental tests have been done with two 3D face datasets FRAV 3D and GAVAB. Therefore, the tests results of the proposed approach are promising because they showed that it is competitive comparable to similar approaches in terms of accuracy, robustness, and flexibility. It achieves a high recognition performance rate of 95.35% for faces with neutral and non-neutral expressions for the identification and 98.36% for the authentification with GAVAB and 100% with some gallery of FRAV 3D datasets.

Road signs classification by ANN for real-time implementation

Traffic safety is an important problem for autonomous vehicles. The development of Traffic Sign Recognition (TSR) dedicated to reducing the number of fatalities and the severity of road accidents is an important and an active research area. Recently, most TSR approaches of machine learning and image processing have achieved advanced performance in traditional natural scenes. However, there exists a limitation on the accuracy in road sign recognition and on time consuming. This paper proposes a real-time algorithm for shape classification of traffic signs and their recognition to provide a driver alert system. The proposed algorithm is mainly composed of two phases: shape classification and content classification. This algorithm takes as input a list of Bounding Boxes generated in a previous work, and will classify them. The traffic signs shape is classified by an artificial neural network (ANN). Traffic signs are classified according to their shape characteristics, as triangular, squared and circular shapes. Combining color and shape information, traffic signs are classified into one of the following classes: danger, information, obligation or prohibition. The classified circular and triangular shapes are passed on to the second ANN in the third phase. These identify the pictogram of the road sign. The output of the second artificial neural network allows the full classification of the road sign. The algorithm proposed is evaluated on a dataset of road signs of a Tunisian database sign.

Automatic crack detection from pavement images using fuzzy thresholding

These last decades have seen the application of automatic inspection in many fields thanks to advanced vision sensors and image analysis methods. However, the difficult nature of pavement images, the small size of defects (cracks) lead to the fact that inspection in this area is done mostly manually. Each year in Tunisia, the operator must view images of thousands of kilometers of roads to detect these degradations. This method is expensive, slow and has a rather subjective result. In this paper, we present a method for automated crack detection. The proposed approach consists of detection of defects. For crack detection, we have applied the method Fuzzy Classification Method (FCM) thresholding. This methodology has been implemented in a Matlab prototype, trained and tested on 330 real pavement images. The results show that this method can detect a crack in pavement images with reasonable accuracy.