The invention discloses an echocardiography value measuring method based on multi-
modal large model workflow, and belongs to the technical field of medical
image processing. According to the method, the problems of low value measurement accuracy and low
report generation efficiency in the traditional cardiac ultrasonic examination are solved. According to the technical scheme, the method comprises the steps of reading an echocardiogram video
stream to obtain a to-be-recognized sequence; preprocessing the data; obtaining section type information through a section classification and
quality control model, wherein the section type information comprises a
cardiac apex four-cavity section, a
cardiac apex two-cavity section and a sternum-side long-axis section; inputting the section type information and the
past medical history of the patient into a large
language model to generate a calling instruction; calling a corresponding visual parameter measurement model according to the instruction to complete
heart structure parameter measurement; and summarizing measured value information, inputting the measured value information into the large
language model, and automatically generating a structured
cardiac ultrasound examination report in combination with measured value parameters and
clinical information. According to the method, multi-model cooperative
processing is realized through multi-
modal information fusion and intelligent scheduling, and the
cardiac ultrasound measurement value accuracy and the diagnostic report normalization are improved.