The invention relates to the technical field of medical monitoring, and discloses an intelligent early warning method and
system for
local anesthetic poisoning, and the early warning method comprises the steps: obtaining multi-
modal data containing the structural features of a patient, physiological
time sequence data and an initial ultrasonic image; based on the ultrasonic image, generating a guide
signal representing a space risk and a depth space
feature vector in parallel; adaptive attention weighting is carried out on the physiological
time sequence data under the guidance of a guidance
signal representing the spatial risk, and an adaptive
time sequence feature vector is generated; and finally, fusing the structural features, the depth space features and the adaptive time sequence features, determining a poisoning
risk level, and performing early warning. By fusing the multi-
modal data and utilizing the space risk analyzed by the ultrasonic image to guide the analysis focus of the
time series data, the method can capture and predict the weak physiological change of the
toxic reaction, thereby overcoming the defects of strong subjectivity and response lagging of the traditional monitoring, and improving the timeliness, accuracy and objectivity of early warning.