卷积神经网络训练及图像处理的方法和系统、计算机设备

Convolutional neural network training and image processing method and system and computer equipment

Abstract

The invention discloses a convolutional neural network training and image processing method, a convolutional neural network training and image processing system and computer equipment. The convolutional neural network training method comprises the following steps: respectively performing image block partition on each of a plurality of images so as to obtain an image block set; generating at least one first image block group according to each image block in the image block set; training the convolutional neural network according to the at least one first image block group; classifying the image blocks in the image block set based on the first convolutional neural network so as to obtain at least one second image block group, wherein the first convolutional neural network is a convolutional neural network trained by the at least one second image block group; extracting the characteristic information of each image block in the image block set based on the first convolutional neural network; and training the first convolutional neural network according to the extracted characteristic information and the at least one second image block group. According to the scheme in the embodiment of the invention, the image blocks in the images without any calibration can be grouped, and non-supervision label calibration is realized.
本申请公开了一种卷积神经网络训练及图像处理的方法和系统、计算机设备,卷积神经网络的训练方法包括:对多个图像中的各图像分别进行图像块划分以得到图像块集合;根据图像块集合中的各图像块生成至少一个第一图像块组;根据至少一个第一图像块组训练卷积神经网络;基于第一卷积神经网络对图像块集合中的各图像块进行分类以得到至少一个第二图像块组,其中,第一卷积神经网络为经至少一个第一图像块组训练后的卷积神经网络;基于第一卷积神经网络提取图像块集合中的各图像块的特征信息;以及根据所提取的特征信息和至少一个第二图像块组训练第一卷积神经网络。采用本申请实施例的方案,可以对无任何标定的图像中的各图像块进行分组,实现无监督标签标定。

Claims

Description

Topics

Download Full PDF Version (Non-Commercial Use)

Patent Citations (0)

    Publication numberPublication dateAssigneeTitle

NO-Patent Citations (0)

    Title

Cited By (0)

    Publication numberPublication dateAssigneeTitle