Tumblelog by Soup.io
Newer posts are loading.
You are at the newest post.
Click here to check if anything new just came in.
1149 2e5a


Deep Convolutional Inverse Graphics Network

Research paper from MIT uses deep learning and neural networks to ‘hallucinate’ a 3D face from a 2D image:

Deep Convolution Inverse Graphics Network (DC-IGN), a model that learns an interpretable representation of images. This representation is disentangled with respect to transformations such as out-of-plane rotations and lighting variations … We propose a training procedure to encourage neurons in the graphics code layer to represent a specific transformation (e.g. pose or light). Given a single input image, our model can generate new images of the same object with variations in pose and lighting. We present qualitative and quantitative results of the model’s efficacy at learning a 3D rendering engine.

More Here

Don't be the product, buy the product!