Traffic Scene Classification

Konzulens:
Szemenyei Márton
Kölső konzuelns:
Continental
Tárgy:
Önálló laboratórium - Szoftverfejlesztés és rendszertervezés specializáció, BSc Info.
Önálló laboratórium - Irányítórendszerek ágazat, BSc Vill.
Hallgatói létszám:
1
Folytatás:
Szakdolgozat / Diplomaterv
Leírás:
Description
Task is to recognize the type of environment from the images of a vehicle mounted camera.
  • Road type: highway, country road, city street
  • Tunnel
  • Weather: Sunny, overcast, rainy, snow
  • Daylight: day / night
  • etc
Goal is to report that status to driver assistance functions. For example, visibility range might be
reduced during heavy rain, so that object detection at higher distances becomes unreliable.
Input to the network shall be a single image, or a video stream of images. Desired output is a set of
predefined categories.
 
Requirements
  • Hands-on experiences with deep learning tools like Keras, Caffe, or similar.
  • Strong programming skills in Python
 
Work can be done as an internship, Bachelor thesis, or project work. Duration 3 - 6 months.