LivePyxel: Accelerating image annotations with a Python-integrated webcam live streaming

The dataset used to train the models presented in the LivePyxel paper: 00_dataset_EM: contains the image/mask pairs of microorganisms taken from a water sample, collected in Toronto, ON, CA, in April 2025. The dataset contains 9 taxonomic groups. 00_snails_aug_dataset_10K: contains the image/mask pairs of snail shells. Each image was generated by superimposing different source images, each representing a morphologically distinct phenotype, to generate 10K image/mask pair datapoints. 00_dataset_snails: contains the unedited, original image/mask pairs of each distinctive morphospecies of snail shell, later edited to generate the augmented dataset 00_snails_aug_dataset_10K Each file is compressed into a .tar file. The structure of each folder is as follows: ./dataset/images ./dataset/masks In addition to these files, a copy of the repository (i.e. the code) with the name 000_Published_Repo_LivePyxel can be also downloaded directly.