The Future: New AI Learns How To See In The Dark

June 6, 2018


These are several demonstrations of a new learning-based artificial intelligence system for processing low-light image that make them appear significantly brighter and more accurately colored than current methods. Some more info while I speculate how I can use this new technique in real-time on trips to the bathroom in the middle of the night. I peed in a towel closet:

Imaging in low light is challenging due to low photon count and low SNR. Short-exposure images suffer from noise, while long exposure can lead to blurry images and is often impractical. A variety of denoising, deblurring, and enhancement techniques have been proposed, but their effectiveness is limited in extreme conditions, such as video-rate imaging at night. To support the development of learning-based pipelines for low-light image processing, we introduce a dataset of raw short-exposure night-time images, with corresponding long-exposure reference images. Using the presented dataset, we develop a pipeline for processing low-light images, based on end-to-end training of a fully-convolutional network. The network operates directly on raw sensor data and replaces much of the traditional image processing pipeline, which tends to perform poorly on such data. We report promising results on the new dataset, analyze factors that affect performance, and highlight opportunities for future work.

Sure being able to recover almost entirely black images is nice, but I want to be able to SEE in the dark. How do I do that? "Night vision goggles." That's expensive, plus only greenvision. I'm thinking maybe I'll just have my eyes replaced with CAT EYES. "You eyes are already beady enough, GW." My girlfriend does say I look like a rat. Thank God she's always had a crush on Master Splinter, am I right? "Can I say no?" Yes but just remember the truth hurts.

Keep going for the video.

Previous Post
Next Post