Key point detection method and apparatus, storage medium and electronic device

A detection method and key point technology, applied in the field of machine learning, can solve the problems of inaccurate positioning of key points, large amount of calculation, time-consuming calculation, etc., and achieve the effect of reducing the amount of calculation, reducing the time-consuming calculation, and accurate positioning

Inactive Publication Date: 2018-06-29
SENSETIME GRP LTD
View PDF4 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] For example, in video-based human body key point detection, most still rely on image-based human body key detection to process continuous video frames, which can easily cause problems such as inaccurate key point positioning, large amount of calculation, and time-consuming calculation.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Key point detection method and apparatus, storage medium and electronic device
  • Key point detection method and apparatus, storage medium and electronic device
  • Key point detection method and apparatus, storage medium and electronic device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0031] figure 2 It is a flow chart showing a key point detection method according to Embodiment 1 of the present invention.

[0032] refer to figure 2 , in step S210, video optical flow data is acquired according to the first video frame and the second video frame included in the video sequence, and the video optical flow data is used to indicate the displacement data of each pixel between the second video frame and the first video frame. Wherein, the second video frame includes at least one video frame that is sequential in sequence and located before the first video frame in the video sequence.

[0033] Among them, the video sequence may include but not limited to live video, recorded video, human-computer interaction video, game video, surveillance video and so on. The first video frame and the second video frame are continuous video frame images in the same video content, and each of the first video frame and the second video frame includes a plurality of pixels and in...

Embodiment 2

[0044] image 3 It is a flow chart showing a key point detection method according to Embodiment 2 of the present invention.

[0045] refer to image 3 , in step S310, video optical flow data is acquired according to the first video frame and the second video frame through a deep neural network for generating video optical flow data.

[0046] In this embodiment, when the key point detection of the target object is performed on the continuous video frames of the video sequence, the continuous second video frame and the first video frame are processed by a deep neural network for generating video optical flow data, thereby obtaining the video The optical flow data indicates the displacement data of each pixel between the second video frame and the first video frame.

[0047] When performing this step, by inputting the second video frame and the first video frame into the trained deep neural network, the video optical flow data between the second video frame output by the deep n...

Embodiment 3

[0066] Figure 4 is a logic block diagram showing a key point detection device according to Embodiment 3 of the present invention.

[0067] refer to Figure 4, the key point detection apparatus of this embodiment includes a first acquisition module 402 and a second acquisition module 404 . The first acquisition module 402 is configured to acquire video optical flow data according to the first video frame and the second video frame included in the video sequence, and the video optical flow data is used to indicate that between the second video frame and the first video frame Displacement data of each pixel, the second video frame includes at least one video frame in the video sequence that is sequential in time sequence and located before the first video frame. The second obtaining module 404 is configured to obtain the first key point of the target object in the first video frame according to the obtained second key point data of the target object in the second video frame a...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

Embodiments of the invention provide a key point detection method and apparatus, a storage medium and an electronic device. The key point detection method comprises the steps of obtaining video optical flow data according to first video frames and second video frames comprised in a video sequence, wherein the video optical flow data is used for indicating displacement data of pixels between the second video frames and the first video frames, and the second video frames include at least one video frame having a continuous time sequence and located before the first video frames in the video sequence; and according to obtained second key point data of target objects in the second video frames and the video optical flow data, obtaining first key point data of target objects in the first videoframes. By adopting the method and the apparatus, the time sequence information of the continuous video frames can be effectively used for accurately locating key points in the continuous video frames.

Description

technical field [0001] Embodiments of the present invention relate to machine learning technology, and in particular to a key point detection method, device, storage medium and electronic equipment. Background technique [0002] Key point detection of objects (such as pedestrians, animals, vehicles, etc.) is an important technology used in applications involving video content analysis and retrieval, and is widely used in robotics, game entertainment, content analysis and recommendation of video websites, etc. At present, the research on object key point detection is divided into image-based key point detection and video-based key point detection. [0003] For example, in video-based human key point detection, most still rely on image-based human key detection to process continuous video frames, which can easily cause problems such as inaccurate key point positioning, large amount of calculation, and time-consuming calculation. Contents of the invention [0004] An embodim...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06N3/04
CPCG06V20/46G06V2201/07G06N3/045
Inventor 张展鹏孙书洋张伟
Owner SENSETIME GRP LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products