Image processing method and device and electronic equipment

Abstract

The invention provides an image processing method and device and electronic equipment, and the method comprises the steps: obtaining a first weight size which is the target weight size of a convolution kernel in a convolution neural network; splitting the first weight size into second weight sizes corresponding to M * N channels, wherein the second weight size is the weight size of a convolution kernel capable of supporting convolution operation by a current hardware facility of the electronic equipment; obtaining first image data corresponding to the corresponding channel number and a secondweight size to perform convolution operation, accumulating obtained convolution results to obtain a convolution result obtained by performing convolution operation on the original image data by usingthe first weight size, the first image data being image data in the original image data, therefore, convolution operation can be performed by using the convolution kernel with any weight size in the convolution neural network model of the current equipment under the condition of not changing hardware design or slightly changing the hardware design, and the image effect output by the convolution neural network is improved.

Description

technical field [0001] The present application relates to the technical field of electronic equipment, in particular to an image processing method, device and electronic equipment. Background technique [0002] As one of the important methods of machine learning, convolutional neural network is widely used in image recognition, object tracking, voice processing and other fields, especially for its excellent performance in processing large images, which has attracted a large number of scholars to engage in research. [0003] At present, in practical application scenarios, the convolutional layer, as an important part of the neural network, participates in more than 90% of the calculation operations of the neural network, and the convolution kernel, as the core part of the convolutional layer, plays a huge role in the neural network model. effect. In the general neural network model, the weight size and stepping of convolution kernels of various sizes exist, from the general ...

AI Technical Summary

Problems solved by technology

In the general neural network model, the weight size and stepping of convolution kernels of various sizes exist, from the general 3×3, 7×7 to 11×11, 15×15, and there may even be weights of the full image size Size, however, in conventional hardware design, the current hardware facilities of electronic equipment can only support a fixed weight size range, such as 1~15, 1~255, etc., which does not support the weight size of any size, so when the neural network When the weight size of the convolution kernel of the model exceeds the weight size of the largest convolution kernel that the current hardware facilities of the electronic device can support, the image processing effect will be poor, and the image processing effect that the neural network model actually wants to achieve cannot be achieved. applications that limit the

Images