Video feature extraction method and device
AssigneeBeijing Bytedance Network Technology Co., Ltd.
InventorsYi He, Lei Li, Cheng Yang, Gen Li, Yitan Li
Life of this patent
Estimating the term…
Abstract
A video feature extraction method and device are provided. The method includes extracting one or more frames from a video object to obtain one or more frames of images; stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images; and determining a video feature according to the image feature of the one or more frames of images. The multi-typed pooling processes include a maximum pooling process, a minimum pooling process and an average pooling process.
Claims15
A video feature extraction method, comprising:
- extracting one or more frames from a video object to obtain one or more frames of images;
- stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images, wherein the multi-typed pooling processes comprise a maximum pooling process, a minimum pooling process and an average pooling process, and wherein the stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images comprises:
- determining a matrix according to the one or more frames of images, stage-by-stage generating a smaller matrix by utilizing the multi-typed pooling processes until the matrix is minimized to be a matrix merely comprising one point, and determining the image feature according to the matrix merely comprising one point;
- and
- determining a video feature according to the image feature of the one or more frames of images.
A video feature extraction device, comprising:
- a memory, configured for storing non-transitory computer readable instructions;
- and
- a processor, configured for executing the non-transitory computer readable instructions such that the processor causes the device to:
- extract one or more frames from a video object to obtain one or more frames of images;
- stage-by-stage process each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images, wherein the multi-typed pooling processes comprise a maximum pooling process, a minimum pooling process and an average pooling process, and wherein stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images comprises:
- determining a matrix according to the one or more frames of images, stage-by-stage generating a smaller matrix by utilizing the multi-typed pooling processes until the matrix is minimized to be a matrix merely comprising one point, and determining the image feature according to the matrix merely comprising one point;
- and
- determine a video feature according to the image feature of the one or more frames of images.
A non-transitory computer readable storage medium, configured for storing non-transitory computer readable instructions, when the non-transitory computer readable instructions are executed by a computer, causing the computer to perform operations of:
- extracting one or more frames from a video object to obtain one or more frames of images;
- stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images, wherein the multi-typed pooling processes comprise a maximum pooling process, a minimum pooling process and an average pooling process, and wherein the stage-by-stage processing each of the one or more frames of images by multi-typed pooling processes to obtain an image feature of the one or more frames of images comprises:
- determining a matrix according to the one or more frames of images, stage-by-stage generating a smaller matrix by utilizing the multi-typed pooling processes until the matrix is minimized to be a matrix merely comprising one point, and determining the image feature according to the matrix merely comprising one point;
- and
- determining a video feature according to the image feature of the one or more frames of images.
Dependencies are inferred from the claim text.