/Patent Tracker
GooglePatent applicationAI & softwareInterest 4/10

Google patents video filter that saves memory by using uneven bit sizes

Google describes a method for cleaning up video images by using a specific type of neural network called a depthwise separable convolutional network. The system saves data space by storing different parts of the network with different levels of detail, keeping important parts precise while simplifying others.

Why it mattersThis technique could help video players use less memory and battery when processing high-quality video streams.

In the filing’s own words

Disclosed are methods, devices, and computer-readable medium for loop filtering using heterogeneously quantized depthwise separable convolutional neural networks. A method includes producing a filtered block of pixels by applying a heterogeneously quantized depthwise separable convolutional neural network (HQDSCNN) to a decoded block of pixels. The HQDSCNN includes a plurality of layers each having components represented using a respective number of bits pre-determined according to a heterogeneous quantization process including identifying a set of key components, setting a minimum number of bits for representing non-key components, and iteratively evaluating candidate numbers of bits for representing components of the set of key components.

What exactly they claim (claim 1)

1. A method comprising: producing a filtered block of pixels by applying a heterogeneously quantized depthwise separable convolutional neural network (HQDSCNN) to a decoded block of pixels, wherein the HQDSCNN includes a plurality of layers each having components represented using a respective number of bits pre-determined according to a heterogeneous quantization process including identifying a set of key components, setting a minimum number of bits for representing non-key components, and iteratively evaluating candidate numbers of bits for representing components of the set of key components.

Facts

Company
Google (Google LLC)
Official title
LOOP FILTER USING HETEROGENOUSLY QUANTIZED DEPTHWISE SEPARABLECONVOLUTIONAL NEURAL NETWORK
Publication number
US20260289247A1
Application number
19570745
Filed
Mar 18, 2026
Published
Sep 24, 2026
Inventors
5
Claims
20

Classified under

G06N 3/0464 Physics & computingG06N 3/0495 Physics & computing

More from Google in AI & software

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.