Bruker:Kaiaw

Fra Robin

Gå til: navigasjon, søk

Master's Thesis

Convolutional networks for road-scene understanding

To make a vehicle drive autonomously, images from color cameras are segmented into seperate regions representing different classes. Pre-trained deep convolutional neural networks is adapted to perform this task by supplementing with road data, and fusing low/high level information from layers of different depth.

Personlige verktøy