# US 2024/0205402 A1 · Grouped feature map quantisation

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2024/0205402 A1
- Document ID: US20240205402A1
- Status: Published application
- Kind: A1
- Application number: 18/554,120
- Filed: 2022-03-11
- Published: 2024-06-20
- Earliest claimed: 2021-04-07
- Assignees: Canon Kabushiki Kaisha
- Applicants: Canon Kabushiki Kaisha
- Inventors: Christopher James Rosewarne; Jonathan GAN
- Web page: https://patentgrep.com/patent/US20240205402A1

## CPC

- G06N 3/063: using electronic means
- H04N 19/124: Quantisation
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/0495: Quantised networks; Sparse networks; Compressed networks
- H04N 19/136: Incoming video signal characteristics or properties
- H04N 19/176: the region being a block, e.g. a macroblock
- H04N 19/20: using video object coding
- H04N 19/88: involving rearrangement of data among different coding units, e.g. shuffling, interleaving, scrambling or permutation of pixel data or permutation of transform coefficient data among different blocks
- G06N 3/048: Activation functions
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

A method of decoding feature maps from encoded data. A parameter related to quantisation is determined. According to the parameter, inverse quantisation is performed for samples decoded from the encoded data to decode the feature maps. For at least a part of the samples decoded from the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

## Claims

### Claim 1

A method of decoding feature maps from encoded data, the method comprising:
determining a parameter related to quantisation; and
performing, according to the parameter, inverse quantisation for samples decoded from the encoded data to decode the feature maps, wherein, for at least a part of the samples decoded from the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

### Claim 2 (depends on 1)

The method according to claim 1, wherein, the samples decoded from the encoded data are arrangeable in a two-dimensional array, and
wherein, samples corresponding to the group of samples are adjacent in the two dimensional array.

### Claim 3 (depends on 1)

The method according to claim 1, further comprising deinterleaving samples corresponding to the group of samples after inverse quantisation is performed.

### Claim 4

An apparatus for decoding feature maps from encoded data, the apparatus comprising:
a determining unit for determining a parameter related to quantisation; and
a quantisation unit for performing, according to the parameter, inverse quantisation for samples decoded from the encoded data to decode the feature maps, wherein, for at least a part of the samples decoded from the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

### Claim 5

A non-transitory computer-readable storage medium which stores a program for executing a method decoding feature maps from encoded data, the method comprising:
determining a parameter related to quantisation; and
performing, according to the parameter, inverse quantisation for samples decoded from the encoded data to the feature maps, wherein, for at least a part of the samples decoded from the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

### Claim 6

A method of encoding feature maps into encoded data, the method comprising:
determining a parameter related to quantisation; and
performing, according to the parameter, quantisation for samples to be encoded into the encoded data to encode the feature maps, wherein, for at least a part of the samples to be encoded into the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

### Claim 7 (depends on 6)

The method according to claim 6, wherein, the samples to be encoded into the encoded data are arrangeable in a two-dimensional array, and
wherein, samples corresponding to the group of samples are adjacent in the two-dimensional array.

### Claim 8 (depends on 6)

The method according to claim 6, further comprising interleaving samples corresponding to the group of samples after quantisation is performed.

### Claim 9

An apparatus for encoding feature maps into encoded data, the apparatus comprising:
a determining unit for determining a parameter related to quantisation; and
a quantisation unit for performing, according to the parameter,
quantisation for samples to be encoded into the encoded data to encode the feature maps, wherein, for at least a part of the samples to be encoded into the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.

### Claim 10

A non-transitory computer-readable storage medium which stores a program for executing a method encoding feature maps into encoded data, the method comprising:
determining a parameter related to quantisation; and
performing, according to the parameter, quantisation for samples to be encoded into the encoded data to encode the feature maps, wherein, for at least a part of the samples to be encoded into the encoded data, one parameter is shared by a group of samples corresponding to a plurality of feature maps.