Personalized Exposure Control Using Adaptive Metering and Reinforcement Learning
This addresses the problem of poor exposure adjustments in mobile photography for users, but it is incremental as it builds on existing metering techniques with personalization.
The paper tackles real-time exposure control for mobile cameras by proposing a reinforcement learning approach that optimizes image quality, fast convergence, and minimal oscillation, showing improved visual quality compared to native camera control in experiments on MIT FiveK and custom datasets.
We propose a reinforcement learning approach for real-time exposure control of a mobile camera that is personalizable. Our approach is based on Markov Decision Process (MDP). In the camera viewfinder or live preview mode, given the current frame, our system predicts the change in exposure so as to optimize the trade-off among image quality, fast convergence, and minimal temporal oscillation. We model the exposure prediction function as a fully convolutional neural network that can be trained through Gaussian policy gradient in an end-to-end fashion. As a result, our system can associate scene semantics with exposure values; it can also be extended to personalize the exposure adjustments for a user and device. We improve the learning performance by incorporating an adaptive metering module that links semantics with exposure. This adaptive metering module generalizes the conventional spot or matrix metering techniques. We validate our system using the MIT FiveK and our own datasets captured using iPhone 7 and Google Pixel. Experimental results show that our system exhibits stable real-time behavior while improving visual quality compared to what is achieved through native camera control.