The application relates to an
endometriosis organoid intelligent monitoring method,
system, device and medium method and
system. The method comprises the following steps: acquiring bright field images of an experimental culture cavity and a reference culture cavity in a microfluidic double-chamber, acquiring
temperature sensing data and
pH sensing data of the two chambers; inputting a lightweight
convolutional neural network for foreground segmentation to obtain
organoid segmentation masks of the two chambers; extracting morphological features based on the segmentation masks to identify invasive phenotypes, and obtaining invasive indexes and invasive
phenotype labels of the two chambers; combining the segmentation masks and the
pH sensing data to calculate a target
perfusion rate; driving a microfluidic
perfusion system to regulate the flow rate, performing temperature
proportional integral differential control according to the
temperature sensing data to generate temperature regulation state data; and integrating the above data to generate an
endometriosis organoid intelligent monitoring result. The method can realize intelligent monitoring and closed-loop regulation of the
culture environment of
endometriosis organoids, and improve the accuracy and experimental efficiency of organoid invasiveness evaluation.