Target segmentation tracking method and device, electronic equipment and storage medium

The embodiment of the invention provides a multi-target segmentation tracking method and device, electronic equipment and a storage medium. The method comprises the steps: firstly, conducting convolutional coding on a to-be-processed image through a convolutional neural network, and constructing a feature pyramid; predicting the category and position of each target according to the feature pyramid, and calculating the centroid coordinate of each target; then carrying out regression on the contour of each target, and calculating coordinates of contour points of each target based on centroid coordinates; generating tracking vectors of the targets according to the coordinates of the contour points; and finally, matching the tracking vector of each target in the continuous frames to obtain a tracking ID of each target. According to the embodiment of the invention, the coordinate of the contour point of each target is predicted based on the centroid coordinates, a target tracking vector is calculated, a complex mu