Title:
ONLINE VERIFICATION METHOD AND SYSTEM FOR REAL-TIME GESTURE DETECTION
Document Type and Number:
WIPO Patent Application WO/2017/177903
Kind Code:
A1
Abstract:
An online verification method for real-time gesture detection, the method comprising the following steps: step 1, an image acquisition module capturing an image within a visual range in real time; step 2, an embedded terminal performing gesture recognition and track monitoring on the acquired image by loading a trained model; step 3, recording a recognition result and responding to the recognition result; step 4, analyzing the correctness of the recognition result and performing re-training to obtain a new model according to a set rule, and verifying the correctness of the new model; and step 5, using the new model to update the pre-trained model. Compared with traditional gesture recognition based on a non-depth camera, the method provides a more accurate on-line model optimization system, thus facilitating the improvement of recognition accuracy.
Inventors:
ZHANG HONGXIN (CN)
CHEN DINGYI (CN)
CHI LIYING (CN)
CHEN DINGYI (CN)
CHI LIYING (CN)
Application Number:
PCT/CN2017/080117
Publication Date:
October 19, 2017
Filing Date:
April 11, 2017
Export Citation:
Assignee:
YUTOU TECH (HANGZHOU) CO LTD (CN)
International Classes:
G06K9/00
Foreign References:
| CN105205436A | 2015-12-30 | |||
| CN103632143A | 2014-03-12 | |||
| US20150309579A1 | 2015-10-29 |
Attorney, Agent or Firm:
SHANGHAI SHENXIN LAW FIRM (CN)
Download PDF: