Previous efforts to restore images have focused on training a neural network on example pairs of noisy and clean images, enabling the artificial intelligence (AI) to learn how to make up the difference. The new AI technique however, developed by researchers from NVIDIA, Aalto University, and MIT, only requires two input images with the noise or grain.
The AI, say the researchers, can remove artifacts, noise, grain, and automatically enhance photos without ever being shown what a noise-free image looks like.
“It is possible to learn to restore signals without ever observing clean ones, at performance sometimes exceeding training using clean exemplars,” say the researchers in a paper on the research. “[The neural network] is on par with state-of-the-art methods that make use of clean examples — using precisely the same training methodology, and often without appreciable drawbacks in training time or performance.”
The researchers trained their system on 50,000 images in the ImageNet validation set using Nvidia Tesla P100 GPUs with the cuDNN-accelerated TensorFlow deep learning framework. To test their system, the researchers validated the neural network on three different datasets.
Potential applications are many, say the researchers. The method can even be used to enhance MRI images, perhaps paving the way to drastically improve medical imaging.
“There are several real-world situations where obtaining clean training data is difficult: low-light photography (e.g., astronomical imaging), physically-based rendering, and magnetic resonance imaging,” they say. “Our proof-of-concept demonstrations point the way to significant potential benefits in these applications by removing the need for potentially strenuous collection of clean data. Of course, there is no free lunch – we cannot learn to pick up features that are not there in the input data – but this applies equally to training with clean targets.”
For more, see “Noise2Noise: Learning Image Restoration without Clean Data.” (PDF)
New Nvidia platform fuses AI, high-performance computing
Nvidia supercomputing platform to ‘revolutionize’ medical imaging
AI more effective than humans at analyzing heart scans
Google acquires AI vision app startup
Food recognition AI turns photos into recipes