TensorFlow implementation of “Simple Does It: Weakly Supervised Instance and Semantic…

“Simple Does It: Weakly Supervised Instance and Semantic Segmentation”, by Khoreva et al. (CVPR 2017)

TensorFlow implementation of “Simple Does It: Weakly Supervised Instance and Semantic Segmentation”, by Khoreva et al. (CVPR 2017)

The repo at https://github.com/philferriere/tfwss contains a TensorFlow implementation of weakly supervised instance segmentation as described in “Simple Does It: Weakly Supervised Instance and Semantic Segmentation”, by Khoreva et al. (CVPR 2017) (https://arxiv.org/abs/1603.07485).

The idea behind weakly supervised segmentation is to train a model using cheap-to-generate label approximations (e.g., bounding boxes) as substitute/guiding labels for computer vision classification tasks that usually require very detailed labels.

This work was done as a contribution to the Monday evening meetup of the Deep Learning Study Group from Silicon Valley Hands On Programming Events (https://www.meetup.com/HandsOnProgrammingEvents) hosted by Mike Bowles. Newcomers and additional contributions (semantic segmentation? Grabcut+?) welcome!

Source: Deep Learning on Medium