Nvidia to Use AI to 'Clean Up' Photos

Researchers are confident that their method will help improve the quality of astronomical and MRI images
11 July 2018   795

Researchers from NVIDIA, MIT and Aalto University showed how to reduce the noise level in photos using AI. The team trained its Noise2Noise system for 50,000 images from the ImageNet suite, using NVIDIA Tesla P100 graphics processors and the TensorFlow framework with cuDNN acceleration.

Usually, neural networks look for the difference between two kinds of photographs: noisy and "clean". The new method does not require the preparation of such pairs, the system only provides shots with different levels of interference for training. It determines how to improve the quality of the image, while not inferior to the old methods of correction.


"Noise" is most often found in MRI images, as well as in astronomical photos. Researchers are confident that their method will help improve the quality of such visualization.

Noise2Noise 2

The scientists presented their work at the International Conference on Machine Learning in Stockholm (ICML).

Nvidia to Open StyleGan Source Code

This machine learning project allows to create of people faces by imitating photographs
11 February 2019   315

NVIDIA has open source code if developments related to the StyleGAN project, which allows generating images of new faces of people by imitating photographs. The system automatically takes into account aspects of the placement of individuals and makes the result indistinguishable from real photos (most of the respondents could not distinguish the original photos from the generated ones). For the synthesis of individuals, a machine learning system based on a generative-competitive neural network (GAN) is used. The code is written in Python using the TensorFlow framework and published under the Creative Commons BY-NC 4.0 license (for non-commercial use only).

Both ready-made trained models and collections of images for self-learning of a neural network are available for download. The basic model was trained on the basis of the Flickr-Faces-HQ (FFHQ) collection, which includes 70,000 high-quality (1024x1024) PNG images of people's faces. At the same time, the system is not tied to persons - as an example, the variants trained on collections of photographs of cars, cats and beds are shown. It requires one or more NVIDIA graphics cards (Tesla V100 GPU recommended), at least 11 GB of RAM, NVIDIA 391.35+ drivers, CUDA 9.0+ tools and the cuDNN 7.3.1 library.

The system allows you to synthesize the image of a new face based on interpolation of features of several faces, combining features characteristic of them, as well as adapting the final image to the required age, gender, hair length, smile character, nose shape, skin color, glasses, face rotation in the photo. The generator considers the image as a collection of styles, automatically separates the characteristic details (freckles, hair, glasses) from common high-level attributes (posture, gender, age changes) and allows you to combine them in an arbitrary form with the definition of the dominant properties through weights.