Skip to content

Latest commit

 

History

History
17 lines (10 loc) · 620 Bytes

README.md

File metadata and controls

17 lines (10 loc) · 620 Bytes

Final project for EE 511 Statistical Learning

By Sandesh Banskota and Chau Huynh

This is a CNN algorithm to determine drivable areas by labeling pixels on imagesfrom the Berkeley Deep Drive dataset https://bdd-data.berkeley.edu/ (github repo: https://github.com/ucbdrive/bdd100k/tree/master/bdd100k)

Input: bdd100k/images Label: bdd100k/drivable_maps/color_labels

References:

U-Net: https://arxiv.org/pdf/1505.04597.pdf Github repo: https://github.com/milesial/Pytorch-UNet/tree/master/unet

FCN: https://arxiv.org/pdf/1411.4038.pdf Github repo: https://github.com/wkentaro/pytorch-fcn/tree/master/torchfcn/models