Inferred veins and lobes on nearly 25K grapevine leaves

An inference dataset on nearly 25K aligned grapevine leaves consisting of 1) segmented rgb images, 2) Euler characteristic transform (ECT) images, and 3) inferred veins and lobes (viz). For each of the above outputs, aligned images are saved in both Cartesian and polar coordinate systems. For the inferred images (viz) segmentation classes are saved in the following true rgb colors: midvein lobe: magenta distal lobe: yellow proximal lobe: cyan midvein: blue lateral veins: green The code to run inference on this data can be found linked below. This dataset is the folder inference_pol_cart_aligned saved to ./outputs by running the following code: Prediction of veins and lobes on nearly 25K grapevine leaves: https://zenodo.org/records/18057298 The model used for inference is linked to below:Segmentation and alignment of grapevine leaf veins and lobes in a rotationally invariant space using the Euler Characteristic Transform: https://zenodo.org/records/18040082 The following shows the history of models leading to this one. The models below are primarily used to predict the petiolar junction (which only requires prediction of the full and primary vein masks), which is used as the foundational input for this inference. Segmentation of RGB leaf masks LEAF_SEGMENTATION: https://zenodo.org/records/16883916 FINAL_MASKS: https://zenodo.org/records/16883403 Segmentation of full vein mask Vein segmentation of pre-segmented grapevine leaves using the radial Euler Characteristic Transform: https://zenodo.org/records/17741713 Segmentation of primary veins Primary vein segmentation in grapevine leaves: https://zenodo.org/records/17881166 Segmentation of midvein Midvein segmentation of grapevine leaves: https://zenodo.org/records/18005383 Gaussian heatmap estimation of petiolar junction coordinate Grapevine leaf petiolar junction detection: https://zenodo.org/records/17922822 Segmentation of midvein, distal, and proximal lobes Grapevine leaf lobe detection model: https://zenodo.org/records/18004938