Xiaomi reveals DeepExposure AI algorithm to improve your phone camera images

The AI algorithm is designed to solve a common issue of over or underexposed photos by employing AI.

  • Published: December 6, 2018 3:15 PM IST

The past few devices from Xiaomi have seen a lot of improvement in the camera department and now the company has plans to improve it further. Xiaomi has introduced its DeepExposure AI algorithm, a new way which will help balance exposure, light and details in your photos.

“The accurate exposure is the key of capturing high-quality photos in computational photography, especially for mobile phones that are limited by sizes of camera modules,” notes a DeepExposure AI research paper. “Inspired by luminosity masks usually applied by professional photographers, in this paper, we develop a novel algorithm for learning local exposures with deep reinforcement adversarial learning.”

The AI algorithm is designed to solve a common issue of over or underexposed photos by employing an AI to fix the exposure of an image (via GSMArena). The DeepExposure balances the exposure by dividing the photo into sub-fragments, and their exposure gets adjusted separately. These different parts are then merged again to form the final image.

The DeepExposure AI algorithm is created in Google’s TensorFlow framework on a Nvidia P40 Tesla GPU. Xiaomi notes that its research team used the images from the MIT-Adobe FiveK dataset.

Watch Video: Xiaomi Poco F1 Armoured Edition First Look

As per research paper, the AI algorithm divides the whole process into image segmentation and ‘action-generation’ stage. During the action-generating stage, we concatenate the input low-resolution image, the sub-image, and the direct fusion image as the state. Then, a policy network is exploited to compute different exposures that are applied on the image locally and globally.

  • Published Date: December 6, 2018 3:15 PM IST