Skip to content

toobamukhtar/Deep_Drive_Image_Segmentation

Repository files navigation

Deep_Drive_Image_Segmentation

Semantic segmentation has been a challenging computer vision problem since a few years. This related work includes a semantic segmentation neural network which combines the strengths of U-Net and residual learning for image segmentation. The network has a similar architecture to that of U-Net. In order to get an insight into the whole working procedure kindly find the attached report.

Please look at FinalReport.docx to check the project details.

Releases

No releases published

Packages

No packages published