We present an accurate and efficient system for traffic sign recognition in a real-world driving scene video. The proposed system uses local structure features to achieve high, illumination-invariant accuracy in detection and recognition. We exploit a property of traffic signs, namely, shared boundary shapes, to enhance the speed and accuracy of the detection step. A multi-level SVM structure i...

We witness many people who face disabilities like being deaf, dumb, blind etc. They a lot of challenges and difficulties trying to interact communicate with others. This paper presents new technique by providing virtual solution without making use any sensors. Histogram Oriented Gradient (HOG) along Artificial Neural Network (ANN) have been implemented. The user makes web camera, which takes in...

In this paper, we propose a two-layer framework of jlow control in high-speed A TM wide area networks. At the upper layer of network level, we classify trafic into two categories, delaysensitive and delay-insensitive. With the advancement of perVC queueing in ATM switch design, we propose peak-rate allocation and therefore no jlow control for delay-sensitive traffic, and hop-by-hop credit-based...

By assuming network trafic to be independent fLom each other, the analysis of network pevformance can be simplified However, the real traflc sources may have some correlation which makes their behavior tends to converge or diverge. This phenomenon has tremendous influence on the congestion control. In this paper, we explain the possible reasons of the correlated behavior, namely, top-down and c...

The purpose of this research is development of an algorithm for hardware implementation for number recognition applying in speed traffic-sign recognition system for car driving assistant. We recognize the speed limit of the speed traffic-sign using hardware oriented extraction algorithm. The numbers are recognized by comparing their feature values with the recognized features. The proposed hard...

This paper presents a background trafic model, ( M , P , S ) , designed for use in wide-area packet-switched network simulators. The approach is to model the aggregate trafSlc generated by a large campus network and then partition the aggregate trafic into substreams, one for each campus-level destination in the backbone network. We use packet traces collected at the network exchange point of a...