Image processing method and device, electronic equipment and computer readable storage medium

Abstract

The embodiment of the invention provides an image processing method and device based on a convolutional network algorithm, electronic equipment and a computer readable storage medium. The method comprises the following steps: determining a filling radius required by each layer of a convolution operation layer; calculating the sum of filling radiuses required by each layer of the convolution operation layer; obtaining a convolutional network algorithm input image; filling the input image according to the sum of the filling radiuses to obtain a filled image; determining an operation range of each layer of the convolution operation layer; performing convolution network algorithm operation on the filled image according to each layer of the convolution operation layer; and obtaining an output image. The invention further provides an image processing device based on the convolutional network algorithm, electronic equipment and a computer readable storage medium. According to the technical scheme, repeated occupation of CPU or hardware operation resources can be reduced, code writing is simplified, and the running speed is increased, so that the economic cost is reduced on the whole, andthe management difficulty is reduced.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a method, device, electronic device and computer-readable storage medium for pre-filling convolutional network algorithms. Background technique [0002] The application of convolutional network technology in artificial intelligence is increasing, especially in the display system industry, for algorithms such as image super-resolution, which can effectively enhance the image quality of the display system. However, in the convolutional network algorithm, it is common to perform convolution filling on the input image or the intermediate image. In a pure hardware system, the filling operation is generally performed directly using hardware; Or the filling operation in the hardware system. figure 1 Shows the general common convolutional network algorithm filling image processing method. However, according to the general convolutional network algorithm filling image p...

AI Technical Summary

Problems solved by technology

However, according to the general convolutional network algorithm filling image processing method, no matter whether the filling operation is performed in the software system or the hardware system, there are problems such as repeated occupation of CPU or hardware computing resources, and the writing of software and hardware codes is relatively complicated.

Images