eliahuhorwitz/DeepSIM
A generative model that manipulates images by editing their primitive representations (edges/segmentations) and reconstructing the image via a learned neural network.
Not currently ranked — collecting fresh signals.
star history
DeepSIM learns to map between a primitive representation of an image (such as edges or segmentation) and the image itself, trained from a single training pair with extensive augmentation including thin-plate-spline transforms. The model enables image manipulation by editing the primitive representation and reconstructing the corresponding image through a conditional GAN architecture, supporting applications like flipping eyebrows, changing object shapes, and converting car types.
Frequently asked
- What is eliahuhorwitz/DeepSIM?
- A generative model that manipulates images by editing their primitive representations (edges/segmentations) and reconstructing the image via a learned neural network.
- Is DeepSIM open source?
- Yes — eliahuhorwitz/DeepSIM is an open-source project tracked on heatdrop.
- What language is DeepSIM written in?
- eliahuhorwitz/DeepSIM is primarily written in Python.
- How popular is DeepSIM?
- eliahuhorwitz/DeepSIM has 425 stars on GitHub.
- Where can I find DeepSIM?
- eliahuhorwitz/DeepSIM is on GitHub at https://github.com/eliahuhorwitz/DeepSIM.