Newborn-painful-expression recognition method based on dual-channel-characteristic deep learning

A deep learning and expression recognition technology, applied in neural learning methods, character and pattern recognition, acquisition/recognition of facial features, etc. The effect of robustness

Active Publication Date: 2017-05-17
NANJING UNIV OF POSTS & TELECOMM
View PDF7 Cites 48 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] To meet the needs of developing an automatic assessment system for neonatal pain, the present invention provides a newborn pain expression recognition method based on dual-channel feature deep learning, which solves the problem of insufficient

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
  • Newborn-painful-expression recognition method based on dual-channel-characteristic deep learning
  • Newborn-painful-expression recognition method based on dual-channel-characteristic deep learning
  • Newborn-painful-expression recognition method based on dual-channel-characteristic deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail:

[0047] The present invention provides a newborn pain expression recognition method based on dual-channel feature deep learning, such as figure 1 As shown, the specific steps are as follows:

[0048] A. Newborn facial images were collected, and professional medical staff classified them into n types of expressions according to the degree of pain, and a newborn facial expression image database was established.

[0049] B, The sample in the neonatal facial expression image library is preprocessed by cropping, alignment, and scale normalization to obtain an image with a size of l×l pixels.

[0050] C, the preprocessed neonatal facial expression image is grayscaled, and its local binary pattern (Local Binary Pattern, LBP) feature map is extracted.

[0051] D, Constructing a dual-channel convolutional neural network for deep learning of the image featu...

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 a newborn-painful-expression recognition method based on dual-channel-characteristic deep learning. The newborn-painful-expression recognition method includes the steps that firstly, newborn facial images are grayed, and a Local Binary Pattern (LBP) specific chromatogram is extracted; secondly, grayscale images of the parallelly-input newborn facial images and the characteristics of two channels of LBP characteristic images of the grayscale images are deeply learned with a dual-channel convolutional neural network; finally, the fusion characteristics of the two channels are subjected to expression classification through a classifier based on a softmax, and expressions are divided into the calmness expression, the crying expression, the mild pain expression and the acute pain expression. According to the newborn-painful-expression recognition method, the grayscale images and the characteristic information of the two channels of the LBP characteristic images of the grayscale images are combined, the expressions such as the calmness expression, the crying expression, the mild pain expression and the acute pain expression can be effectively recognized, the quite-good robustness of the illumination problem, the noise problem and the shielding problem of the newborn facial images is achieved, and a new method and way are provided for developing a newborn-painful-expression recognition system.

Description

technical field [0001] The invention relates to a newborn pain expression recognition method based on dual-channel feature deep learning, which belongs to the field of image processing and emotion recognition. Background technique [0002] Pain is a common uncomfortable symptom of the human body. It not only makes people suffer, but also brings a series of physiological and psychological adverse effects. Studies have shown that newborns have the ability to feel pain after birth, and can transmit, perceive, respond to and even remember harmful stimuli. Various examinations and treatments received from birth will bring pain and stimulation to newborns. Painful stimuli can cause systemic reactions in the body, such as changes in breathing and immunity, and instability of cardiovascular function; this pain may also lead to long-term effects such as developmental delay in newborns, permanent damage to the central nervous system, and emotional disorders. Early repeated operant 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): G06K9/00G06N3/08G06N3/04
CPCG06N3/08G06V40/174G06N3/045
Inventor 朱金朵卢官明李晓南闫静杰李海波
Owner NANJING UNIV OF POSTS & TELECOMM
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