# US 2025/0095104 A1 · Image signal processor and image signal processing method > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2025/0095104 A1 - Document ID: US20250095104A1 - Status: Published application - Kind: A1 - Application number: 18/801,852 - Filed: 2024-08-13 - Published: 2025-03-20 - Earliest claimed: 2023-09-15 - Assignees: SK hynix Inc. - Applicants: SK hynix Inc. - Inventors: Yuuki ADACHI - Web page: https://patentgrep.com/patent/US20250095104A1 ## CPC - G06T 5/20: using local operators - G06T 7/13: Edge detection - G06T 3/40: Scaling of whole images or parts thereof, e.g. expanding or contracting - G06T 5/00: Image enhancement or restoration - G06T 7/181: involving edge growing; involving edge linking - H04N 25/68: applied to defects - H04N 25/683: by defect estimation performed on the scene signal, e.g. real time or on the fly detection - G06T 2207/10024: Color image - H04N 25/13: characterised by the spectral characteristics of the filter elements - H04N 25/611: Correction of chromatic aberration - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION ## Abstract An image signal processor includes a directional component extractor configured to extract directional components of a target kernel by performing a convolution operation between the target kernel including a target pixel and each of a plurality of directional kernels, an interpolation kernel determiner configured to determine an interpolation kernel based on the directional components, and a pixel interpolator configured to interpolate the target pixel using data included in the interpolation kernel. ## Claims ### Claim 1 An image signal processor comprising: a directional component extractor configured to extract directional components of a target kernel by performing a convolution operation between the target kernel including a target pixel and each of a plurality of directional kernels; an interpolation kernel determiner configured to determine an interpolation kernel based on the directional components; and a pixel interpolator configured to interpolate the target pixel using data included in the interpolation kernel. ### Claim 2 (depends on 1) The image signal processor according to claim 1, wherein the interpolation kernel determiner is configured to obtain convolution resultant values by performing a convolution operation between each of a plurality of cost kernels and the directional components, and determine the interpolation kernel based on the convolution resultant values. ### Claim 3 (depends on 2) The image signal processor according to claim 2, wherein each of the plurality of cost kernels includes a preset weight. ### Claim 4 (depends on 2) The image signal processor according to claim 2, wherein the convolution resultant values include result values obtained by performing a point-wise convolution operation in which the directional components are matched to channel and each of the plurality of cost kernels is matched to a unit kernel. ### Claim 5 (depends on 2) The image signal processor according to claim 2, wherein the interpolation kernel includes an interpolation kernel corresponding to a cost kernel that derives a maximum or minimum value from among the convolution resultant values. ### Claim 6 (depends on 1) The image signal processor according to claim 1, wherein the directional component extractor is configured to extract the directional components by using pixel data of each pixel having a first color filter from among pixels included in the target kernel. ### Claim 7 (depends on 6) The image signal processor according to claim 6, wherein the directional component extractor is configured to determine, as a particular value, pixel data of each pixel having a color filter different from the first color filter from among pixels included in the target kernel. ### Claim 8 (depends on 6) The image signal processor according to claim 6, wherein the directional component extractor is configured to determine, as a particular value, data corresponding to a position of each pixel having a color filter different from the first color filter from among data included in the plurality of directional kernels. ### Claim 9 (depends on 1) The image signal processor according to claim 1, wherein the target kernel has the same size as the largest directional kernel from among the plurality of directional kernels. ### Claim 10 (depends on 1) The image signal processor according to claim 1, wherein the pixel interpolator is configured to: match, to the target pixel, data located at a center of the interpolation kernel; and determine, as pixel data of the target pixel, a convolution resultant value between the target kernel and the interpolation kernel. ### Claim 11 An image signal processor comprising: a directional component extractor configured to extract directional components of a target kernel by performing a first convolution operation between the target kernel and each of a plurality of first kernels, the target kernel including a target pixel; an interpolation kernel determiner configured to obtain convolution resultant values by performing a second convolution operation between each of a plurality of second kernels and the directional components, and determine an interpolation kernel based on the convolution resultant values; and a pixel interpolator configured to interpolate the target pixel using data included in the interpolation kernel. ### Claim 12 (depends on 11) The image signal processor according to claim 11, wherein the first convolution operation is performed in a state in which the target pixel is matched to data located at a center of each of the plurality of first kernels. ### Claim 13 (depends on 11) The image signal processor according to claim 11, wherein the second convolution operation is a point-wise convolution operation in which the directional components are matched to a channel and each of the plurality of second kernels is matched to a unit kernel. ### Claim 14 (depends on 11) The image signal processor according to claim 11, wherein the interpolation kernel refers to an interpolation kernel corresponding to a kernel that derives a maximum or minimum value among the convolution resultant values obtained by the second convolution operation, from among kernels included in the plurality of second kernels. ### Claim 15 An image signal processing method comprising: extracting first directional components of a first color kernel by performing a convolution operation between the first color kernel and each of a plurality of directional kernels, the first color kernel including a first target pixel; determining a first interpolation kernel based on the first directional components; and interpolating the first target pixel using data included in the first interpolation kernel. ### Claim 16 (depends on 15) The image signal processing method according to claim 15, further comprising: generating an interpolated target kernel by mixing an interpolated first color kernel with a second color kernel, the interpolated first color kernel including an interpolated first target pixel. ### Claim 17 (depends on 16) The image signal processing method according to claim 16, further comprising: extracting second directional components of the interpolated target kernel by performing a convolution operation between each of the directional kernels and the interpolated target kernel including a second target pixel; determining a second interpolation kernel based on the second directional components; and interpolating the second target pixel using data included in the second interpolation kernel. ### Claim 18 (depends on 15) The image signal processing method according to claim 15, wherein the determining the first interpolation kernel includes obtaining convolution resultant values by performing a convolution operation between each of a plurality of cost kernels and the first directional components, and determining the first interpolation kernel based on the convolution resultant values. ### Claim 19 (depends on 18) The image signal processing method according to claim 18, wherein the determining the first interpolation kernel further includes multiplying each of the plurality of cost kernels by a preset weight. ### Claim 20 (depends on 18) The image signal processing method according to claim 18, wherein the obtaining the convolution resultant values includes performing a point-wise convolution operation by matching the first directional components to channel and matching each of the plurality of cost kernels to a unit kernel.