An improved method for improving stability of video face detection

A face detection and stability technology, applied in the field of video face detection, can solve the problems of large differences in face positions, stiff face position labeling frames, and lagging face positions

Inactive Publication Date: 2019-05-10
广州二元科技有限公司
View PDF1 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when the face detection result obtained by using this method moves rapidly in a short period of time (for example, within 1 second), the position of the face (the label frame of the face position) will lag behind, that is, the proces

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
  • An improved method for improving stability of video face detection
  • An improved method for improving stability of video face detection
  • An improved method for improving stability of video face detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0082] Such as figure 1 As shown, a method for improved video face detection to improve stability is characterized in that it comprises the following steps:

[0083] (1) Directly input a video file containing multiple frames, or collect real-time video through a hardware device camera as input; the input video must be valid and the data type and size of the video image matrix of each frame should be consistent (same within an input video);

[0084] (2) Obtain the frame rate of the current video or the frame rate of the video captured by the camera; usually this frame rate can be returned by calling the relevant API, and the frame rate is recorded in FPS; if the system cannot obtain the FPS or obtains an error, then use the default frame rate value, let

[0085] FPS=30

[0086] And define an array container ARRAY to store a certain number T of face detection results, and the value of T is:

[0087] T = FPS

[0088] That is to say, this embodiment saves the face detection r...

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

The invention discloses an improved method for improving stability by video face detection. The method comprises the following steps: (1) inputting a video; (2) acquiring the frame rate of the currentvideo or the frame rate of the video acquired by the camera; (3) decomposing the video by taking a single frame as a unit to form a digital image matrix of which each input is one frame; (4) inputting one frame of digital image matrix into the face detector, calculating and returning a face detection result, and storing the face detection result into the ARRAY; (5) carrying out result associationcalculation between upper and lower frames on face detection in the video by using simple mathematical operation; carrying out weight offset calculation on a plurality of face similar frames, and averaging the weight offset calculation and the current frame of face frame; (6) adding the current frame of face detection result and the calculated face position marking box into the array; and (7) inputting the next frame of image matrix, and executing the process from the step 5 until the video frame processing is finished. The method has the advantages of simple algorithm, high detection rate and wide applicability.

Description

technical field [0001] The invention relates to the technical field of video face detection, in particular to an improved method for improving stability of video face detection. Background technique [0002] Face detection and face recognition have achieved unprecedented development in the past 20 years. Face detection through Harbin features and integral images, face detection through color template matching, and the current popular neural network structure. Face detection, its accuracy and stability are gradually improving. [0003] Video face detection is a popular subject, and more and more scenarios need to use this technology. Such as digital monitoring system, Selfie system in intelligent mobile equipment, living body recognition system, etc. It is of great significance to ensure the stability of the output face position in most video face detection. In 10 consecutive video frames, when the position of the input face remains basically unchanged, the obtained face p...

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
IPC IPC(8): G06K9/00
Inventor 容李庆袁亚荣黄东生
Owner 广州二元科技有限公司
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