Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. Intra Picture Prediction for Video Coding with Neural Networks
 Bilgin, A. ; University of Arizona; Brandeis University; Microsoft Research; IEEE Signal Processing Society: Data Compression Conference, DCC 2019. Proceedings : Snowbird, Utah, USA, 2629 March 2019 Piscataway, NJ: IEEE, 2019 ISBN: 9781728106571 ISBN: 9781728106588 pp.448457 
 Data Compression Conference (DCC) <2019, Snowbird/Utah> 

 English 
 Conference Paper 
 Fraunhofer HHI () 
Abstract
We train a neural network to perform intra picture prediction for block based video coding. Our network has multiple prediction modes which coadapt during training to minimize a loss function. By applying the l1norm and a sigmoidfunction to the prediction residual in the DCT domain, our loss function reflects properties of the residual quantization and coding stages present in the typical hybrid video coding architecture. We simplify the resulting predictors by pruning them in the frequency domain, thus greatly reducing the number of multiplications otherwise needed for the dense matrixvector multiplications. Also, by quantizing the network weights and using fixed point arithmetic, we allow for a hardware friendly implementation. We demonstrate significant coding gains over state of the art intra prediction.