Human-computer interaction method based on artificial intelligence
A technology of human-computer interaction and artificial intelligence, applied in the input/output of user/computer interaction, mechanical mode conversion, computer components, etc., can solve the problems of slow recognition speed, difficult to distinguish, poor user experience, etc., and achieve recognition speed Fast, high accuracy, and high recognition frame rate
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0073] A total of 15G video training sets were created by crawling and combining open source gesture recognition videos. The dynamic gestures in the video include zooming, panning, clicking, grabbing, and rotating.
[0074] The recognition model is built through the TensorFlow deep learning engine. The recognition model includes a spatial channel sub-model and a time channel sub-model, both of which are I3D models. Its structure is as follows figure 2 shown.
[0075] Process the video clips in the video training set, use OpenCV to extract video clips frame by frame to obtain video frame pictures, and use the Farnback method to process video clips to obtain optical flow estimation, use frame pictures to train the spatial channel sub-model, a total of 6000 training steps , the error tends to 0 after training, and the training error change curve is as follows Figure 5 As shown; the optical flow estimation is used to train the time channel sub-model, a total of 9000 training st...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com