Research Project for CprE 575, Computational Perception
An automatic traffic sign detection system can assist drivers on better understanding their surrounding situations through reading the information from every traffic sign the vehicle has previously encountered. It can help reduce the traffic accident rates by helping drivers remember useful information such as speed limits, warnings for possible dangers, and more, therefore improving driving quality and safety. This system can also be utilized in assisting drivers with disabilities, developing autonomous vehicles, and more.
Nowadays, Traffic Sign Detection is considered a well-researched topic in Computational Perception and Image Processing. Many algorithms are used or developed for this purpose. Some of the popular ones include Support Vector Machines (SVM), Histogram of Gradient (HOG), convolutional neural network [1], Joint Transform Correlator [3], and Color Segmentation [5]. In this project, I will try out some basic image processing approaches using MATLAB and OpenCV, as well as other higher-level machine learning algorithms to improve the accuracy and efficiency of the detection.
The initial project proposal, the final write-up, and the poster used during presentation is uploaded to this repository as they summarize the final stage of the project.