US Patent Application 18336735. DECODING AND ENCODING OF NEURAL-NETWORK-BASED BITSTREAMS simplified abstract
Contents
DECODING AND ENCODING OF NEURAL-NETWORK-BASED BITSTREAMS
Organization Name
Inventor(s)
Elena Alexandrovna Alshina of Munich (DE)
DECODING AND ENCODING OF NEURAL-NETWORK-BASED BITSTREAMS - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 18336735 Titled 'DECODING AND ENCODING OF NEURAL-NETWORK-BASED BITSTREAMS'
Simplified Explanation
The abstract describes a method for decoding and encoding pictures using neural networks. The picture is represented by a set of samples obtained from a bitstream. The picture is reconstructed by processing the input set and generating output subsets. The input set is divided into smaller subsets and each subset is processed by a neural network. The neural network takes multiple samples from an input subset and generates one sample for an output subset. The reconstructed picture is obtained by combining the output subsets. The method involves using smaller input subsets than the required size for the corresponding output subsets after processing by the neural network layers.
Original Abstract Submitted
For picture decoding and encoding of neural-network-based bitstreams, a picture is represented by an input set of samples which is obtained from the bitstream. The picture is reconstructed from output subsets, which are generated as a result of processing the input set L. The input set is divided into multiple input subsets Li. The input subsets are each subject to processing with a neural network having one or more layers. The neural network uses as input multiple samples of an input subset and generates one sample of an output subset. By combining the output subsets, the picture is reconstructed. In particular, the size of at least one input subset is smaller than a size that is required to obtain the size of the respective output subset, after processing by the one or more layers of the neural network.
