Computer vision based evaluation method for general movement of baby

Abstract

The invention provides a computer vision based evaluation method for general movement of a baby. The method comprises the steps of reading an acquired video through a computer; preprocessing the image by fuzzification and channel separation; calculating light flow of adjacent frames; extracting the baby body area; obtaining the movement light flow of the baby body area; displaying the baby movement in the images at all frames through one image; building an u-v right angle distribution map; counting the number of pixels in each part; counting all frames in the same area to obtain a speed counting histogram under different direction and size of all frames; evaluating the general movement of the baby according to the speed distribution in the histogram. The method is applied to complexity, variation and symmetry evaluation detection for general movement of the baby under a scene that a camera is fixed, and has the advantages of being high in generality, high in accuracy and fast to respond.

Description

technical field [0001] The invention relates to a computer vision signal collection and processing technology, in particular to a method for evaluating baby's whole-body motion through computer-collected baby motion video. Background technique [0002] With the improvement of medical level, the survival rate of high-risk neonates (refers to infants who are harmed by high-risk factors such as premature birth, low birth weight, asphyxia, hypoxic-ischemic encephalopathy, and intracranial hemorrhage during the fetal period, delivery period and neonatal period) continues to increase. increased, but this did not reduce the incidence of neurodevelopmental disorders such as cerebral palsy. According to the data, the incidence rate of cerebral palsy in my country is 2 / 1000, and neurodevelopmental disorders have seriously affected the quality of life of patients and brought heavy burdens to families and society. UNICEF proposed in 2000: "The struggle to save children's lives and the ...

AI Technical Summary

Problems solved by technology

Both the acceleration sensor and the electromagnetic tracking device have their disadvantages: the measurement results of the electromagnetic tracking device are greatly affected by the surrounding electromagnetic field, such as EMG, EEG, and metal objects, which are commonly used in clinical practice, will affect the measurement results of the electromagnetic tracking device; Accelerometers will drift during long-term recording; these methods will place markers or sensing points on the baby's body, which may affect the baby's normal activities

At present, the research on the objective evaluation of infant GMs based on video analysis has achieved certain experimental results, but these methods have certain limitations. The features extracted by most methods are very simple, or cannot reflect the direction and size of movement, or only based on the body The trajectory analysis of several points of the baby's movement only utilizes little movement information in the video, and does not analyze the movement from the whole body movement, while the baby's GMs assessment requires overall perception, so how to comprehensively use the baby's body in video analysis All the movement information (including all parts of the body, as well as the direction and size of the movement), and expressing the movement of the body as a whole are still the current problems

Images