Seal Integrity in Sterile Barrier Packaging

Dataset containing 1200 images of Sterile Packaging of size (3008, 4110, 3), where 420 images contain defects and associated mask files. This dataset was used to support publication: "Automatic Inspecting of Seal Integrity in Sterile Barrier Packaging: a Deep Learning Approach". Trained models against this dataset can be found on: - No Augmentation: 10.5281/zenodo.7858060 - Augmentation: 10.5281/zenodo.7858003 File descriptions: - DS1.zip: Images and masks for the full dataset. - Full-50.zip: Information for sub-split "Full-50". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-50.zip: Information for sub-split "Partial-50". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-40.zip: Information for sub-split "Partial-40". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-30.zip: Information for sub-split "Partial-30". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-20.zip: Information for sub-split "Partial-20". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-10.zip: Information for sub-split "Partial-10". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-05.zip: Information for sub-split "Partial-05". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv - Partial-01.zip: Information for sub-split "Partial-01". Files: ds_info.json, Tesnting.csv, TrainingXVal.csv