SpinX pipelines for automated label generation. (a) Conventional image processing pipeline to segment chromosomes. Pipeline includes using a median filter to reduce surrounding noise while preserving information of edges; performing Otsu’s thresholding to create a binary image; removing incomplete objects located at the image boundary; extracting boundary pixel information of the metaphase plate and applying a Savitzky-Golay filter to smoothen the boundary. (b) Conventional image processing pipeline to segment SiR-Tubulin-labeled spindle images. Pipeline includes using a median filter to reduce surrounding noise while preserving information of edges; performing Otsu’s thresholding to create a binary image; calculating the binary convex hull image; extracting boundary pixel information of the spindle; applying a Savitzky-Golay signal processing filter for smoothing; and utilizing an ellipse fit to obtain the final boundary information. (c) Conventional image processing pipeline (non-AI-based image processing techniques) to segment the mCherry-Tubulin-labeled spindle images. Pipeline includes applying a Gaussian filter to reduce surrounding noise while preserving information of edges; calculating the image gradient; using the segmentation mask of the metaphase plate to initiate the inverse snake; extracting boundary pixel information of the spindle and utilizing an ellipse fit to obtain the final boundary information. Scale bars: 10 µm.