High dynamic range (HDR) UHD-TVs are being rapidly deployed in consumer markets, offering a highly realistic experience to customers. However, these HDR UHD-TVs still need to handle the legacy low resolution (LR) video of standard dynamic range (SDR). In this paper, we propose a convolutional neural network based structure for the joint learning of super-resolution and inverse tone-mapping, which can be used for converting LR-SDR legacy video to high resolution (HR) HDR video. Our proposed structure is designed to perform three tasks: (i) SDR-to-HDR conversion of LR images, (ii) super-resolution of LR-SDR images to HR-SDR images and (iii) joint conversion from LR-SDR to HR-HDR images. We show the effectiveness of our proposed joint learning CNN architecture with extensive experiments.
CITATION STYLE
Kim, S. Y., & Kim, M. (2019). A Multi-purpose Convolutional Neural Network for Simultaneous Super-Resolution and High Dynamic Range Image Reconstruction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11363 LNCS, pp. 379–394). Springer Verlag. https://doi.org/10.1007/978-3-030-20893-6_24
Mendeley helps you to discover research relevant for your work.