Image processing method and device and computer storage medium
By decomposing the 3D convolution model into a 2D cascade convolution model and applying feature rearrangement rules, the problem in the existing technology that the 3D convolution model cannot be realized by the 2D convolution model is solved, and multi-frame images or videos are realized. Convolution operation, image classification, action recognition and other functions.
CN111368941AActive Publication Date: 2020-07-03ZHEJIANG DAHUA TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ZHEJIANG DAHUA TECH CO LTD
- Filing Date
- 2020-04-10
- Publication Date
- 2020-07-03
AI Technical Summary
⚠Technical Problem
There is a problem in the existing technology that image processing implemented using 3D convolution models cannot be implemented through 2D convolution models.
⚗Method used
The 3D convolution model is decomposed into a pseudo 3D cascade convolution model and converted into a 2D cascade convolution model. The parameters are mapped through feature rearrangement rules, and the 2D spatial and temporal convolution model is used to perform convolution operations on the image.
🎯Benefits of technology
It implements an image processing method that implements a 3D convolution model through 2D convolution simulation, which can process multiple frames of images or videos and complete tasks such as image classification and action recognition.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses an image processing method and device and a storage medium. The method comprises the steps of obtaining a to-be-simulated 3D convolution model and training data; decomposing the 3D convolution model into cascading of a 3D space convolution model and a 3D time convolution model to obtain a pseudo 3D cascading convolution model; training a pseudo 3D cascade convolution modelby using the training data, and obtaining parameters of a 3D spatial convolution model and a 3D time convolution model; converting the 3D space convolution model and the 3D time convolution model intoa 2D space convolution model and a 2D time convolution model; setting a feature rearrangement rule for the 2D spatial convolution model and the 2D time convolution model; mapping model parameters ofthe 3D spatial convolution model and the 3D time convolution model into parameters of a 2D spatial convolution model and a 2D time convolution model to obtain a 2D cascaded convolution model; and performing convolution operation on the image by using the 2D spatial convolution model and the 2D time convolution model. By means of the mode, image processing conducted through 3D convolution operationcan be achieved through the 2D convolution model.
Need to check novelty before this filing date? Find Prior Art