18479507. CONFIGURABLE POSITIONS FOR AUXILIARY INFORMATION INPUT INTO A PICTURE DATA PROCESSING NEURAL NETWORK simplified abstract (HUAWEI TECHNOLOGIES CO., LTD.)

CONFIGURABLE POSITIONS FOR AUXILIARY INFORMATION INPUT INTO A PICTURE DATA PROCESSING NEURAL NETWORK

Organization Name

HUAWEI TECHNOLOGIES CO., LTD.

Inventor(s)

Timofey Mikhailovich Solovyev of Munich (DE)

Biao Wang of Shenzhen (CN)

Elena Alexandrovna Alshina of Munich (DE)

Han Gao of Shenzhen (CN)

Panqi Jia of Munich (DE)

Esin Koyuncu of Munich (DE)

Alexander Alexandrovich Karabutov of Munich (DE)

Mikhail Vyacheslavovich Sosulnikov of Munich (DE)

Semih Esenlik of Munich (DE)

Sergey Yurievich Ikonin of Moscow (RU)

CONFIGURABLE POSITIONS FOR AUXILIARY INFORMATION INPUT INTO A PICTURE DATA PROCESSING NEURAL NETWORK - A simplified explanation of the abstract

This abstract first appeared for US patent application 18479507 titled 'CONFIGURABLE POSITIONS FOR AUXILIARY INFORMATION INPUT INTO A PICTURE DATA PROCESSING NEURAL NETWORK

Simplified Explanation

This patent application describes an application that uses a neural network with multiple layers to process picture data or picture feature data. The application is specifically designed for artificial intelligence (AI)-based video or picture compression technologies, particularly neural network-based video compression technologies.

Key points of the patent/application:

• The application focuses on processing picture data or picture feature data using a neural network with two or more layers.
• The neural network's configuration can be adjusted by selecting the position within the network where auxiliary information can be entered for processing.
• The selection of the position for auxiliary information is based on a gathering condition, which determines if a prerequisite is fulfilled.
• The flexibility in neural network configurability allows for better performance in terms of rate and/or disclosure.

Potential applications of this technology:

• Video compression: The technology can be used to improve video compression algorithms by utilizing neural networks with multiple layers.
• Picture compression: Similarly, the technology can be applied to picture compression algorithms to enhance compression efficiency.
• Artificial intelligence: The application's use of neural networks aligns with the broader field of artificial intelligence, allowing for advancements in AI-based technologies.

Problems solved by this technology:

• Improved compression efficiency: By utilizing neural networks with multiple layers, the technology aims to enhance the efficiency of video and picture compression algorithms.
• Increased configurability: The ability to select the position for auxiliary information within the neural network provides increased flexibility in configuring the network, potentially leading to better performance.

Benefits of this technology:

• Enhanced compression performance: The use of neural networks with multiple layers can improve the compression performance of video and picture data.
• Increased flexibility: The selectable position for auxiliary information allows for greater configurability, enabling customization based on specific requirements.
• Potential for better rate and disclosure: The increased flexibility and configurability of the neural network can lead to improved compression rates and disclosure of information.

Original Abstract Submitted

This application provides methods and apparatuses for processing of picture data or picture feature data using a neural network with two or more layers. The present disclosure may be applied in the field of artificial intelligence (AI)-based video or picture compression technologies, and in particular, to the field of neural network-based video compression technologies. According to some embodiments, position within the neural network, at which auxiliary information may be entered for processing is selectable based on a gathering condition. The gathering condition may assess whether some prerequisite is fulfilled. Some of the advantages may include better performance in terms of rate and/or disclosure due to the effect of increased flexibility in neural network configurability.