Vehicle Detection and Tracking -- 3

The project involves Detection and Tracking of on road vehicles with following details: a. The input data comes from a video camera of 2MP and at a rate of 30fps. b. Use Artificial Intelligence techniques such as Deep Learning (CNN, etc) to detect vehicle c. Use online datasets for training and validation d. The delay between an object seen by the camera and the output by the software should not be more than 500 milliseconds e. Accuracy of vehicle detection expected to be +/- 2 cm f. Training & Validation accuracy to be above 98% g. Training will need GPU and might be tested on a cloud based framework such as AWS or even intel but application should be able to run on a Macbook machine (i7 Core, 2.2 GHz, 16 GB RAM) h. It is desirable to deploy the application on a real time embedded platform Deliverables: a. Software code including training and application both b. Hardware and software specifications c. Sources of datasets used for training and validation More details can be clarified and discussed after acceptance of the project.