Representing Camera Response Function by a Single Latent Variable and Fully Connected Neural Network

Modelling the mapping from scene irradiance to image intensity is essential for many com-puter vision tasks. Such mapping is known as the camera response. Most digital camerasuse a nonlinear function to map irradiance, as measured by the sensor to an image inten-sity used to record the photograph. Modelling of the response is necessary for the nonlinearcalibration. In this paper, a new high-performance camera response model that uses a singlelatent variable and fully connected neural network is proposed. The model is produced usingunsupervised learning with an autoencoder on real-world (example) camera responses. Neuralarchitecture searching is then used to find the optimal neural network architecture. A latent dis-tribution learning approach was introduced to constrain the latent distribution. The proposedmodel achieved state-of-the-art CRF representation accuracy in a number of benchmark tests,but is over twice as fast as the best current models when performing the maximum likelihoodestimation during camera response calibration due to the simple yet efficient model representation.