This is my attempt of a past Kaggle competition Lyft 3D Object Detection for Autonomous Vehicles, using the concept of frustum-pointnet to solve this past Kaggle competition.

This is an example of the classifed results:

Result

This work is still in progress. The result does not seem too bad if …

I did this project partly for Udacity's Self-Driving car degree and was lazy to train and retraing a DNN model. It ends up that this solution is not so lazy, but I did learn some details and traps of using a HSV or HSL color space.

I posted the writing-up …

The aim of this project is to implement a car that drives automatically in a simulation world. The car should be able to switch the lane smoothly when necessary. I think this is the most fun and challenging project of Udacity Self-Driving Car Nanodegree.

The code repo and the explaination …

This is a project with Udacity's Self-Driving Car Nanodegree. The aim is to use deep learning approach[1] to clone the behaviour of a car in the simulator made by Udacity. To get a good data set, I have to "drive" well in the simulator, and use this dataset to …