The invention relates to the crossing field of medical
image processing and
artificial intelligence technology, in particular to a
bone marrow cell morphology
artificial intelligence analysis method and
system, and the method comprises the steps: obtaining a multi-view high-resolution image of a
bone marrow smear and / or a
peripheral blood smear under the
oil immersion magnification, and carrying out the preprocessing; carrying out
red blood cell coarse segmentation and fine classification through a
cascade deep network, and identifying target-shaped, broken, teardrop-shaped and elliptical red blood cells; correcting broken
red blood cell counting by adopting a fragment aggregation
correction algorithm; and calculating an MCI value in combination with blood routine
test data, and carrying out weighted fusion on the MCI value and the abnormal
red blood cell proportion to generate a
thalassemia screening comprehensive
score. Therefore, the problems of low manual identification efficiency, insufficient abnormal red
blood cell identification accuracy, large broken red
blood cell counting deviation, difficulty in distinguishing multiple types of abnormal red blood cells, lack of organic combination with
clinical diagnosis indexes and the like are solved, and the efficiency and accuracy of
morphological analysis of
bone marrow cells are improved; and a reliable basis is provided for diagnosis of diseases such as
thalassemia.