Image processing method and device using line input and output
AssigneeKorea Advanced Institute of Science and Technology
InventorsMun Churl Kim, Yong Woo Kim, Jae Seok Choi
Life of this patent
Estimating the term…
Abstract
Disclosed are an image processing method and device using a line-wise operation. The image processing device, according to one embodiment, comprises: a receiver for receiving two-dimensional pixel information; at least one first line buffer for outputting the two-dimensional pixel information into a line-wise data line; an operator for performing a convolution operation on the basis of the data line data; and at least one second line buffer for outputting the data line on which the operation has been performed.
Claims26
An image processing device comprising:
- a receiver configured to receive two-dimensional (2D) pixel information;
- at least one first line buffer configured to output the 2D pixel information in a line-wise data line;
- an operator configured to perform a convolution operation based on the 2D pixel information of the data line;
- and
- at least one second line buffer configured to output a data line obtained by performing the convolution operation
- wherein the operator comprises:
- a first convolution operator a first convolution operator configured to generate at least one feature map by performing a convolution operation for the data in the data line,
- a feature map processor configured to store the at least one feature map as at least one line-wise feature map, and process the stored at least one line-wise feature map to be output in a two or higher dimensional form;
- and
- a second convolution operator configured to perform a convolution operation for the at least one feature map in the two or higher dimensional form.
An image processing method comprising:
- receiving two-dimensional (2D) pixel information;
- outputting the 2D pixel information in at least one line-wise data line;
- performing a convolution operation based on the 2D pixel information of the data line;
- and
- outputting a data line obtained by performing the convolution operation,
- wherein the performing comprises:
- generating at least one feature map by performing a first convolution operation for the data in the data line,
- storing the at least one feature map as at least one line-wise feature map, and processing the stored at least one line-wise feature map to be output in a two or higher dimensional form;
- and
- performing a second convolution operation for the at least one feature map in the two or higher dimensional form.
Dependencies are inferred from the claim text.