VIA is an image tool for visualizing and editiing object detection datasets. It allows the user to determine quality of annotations to verify the integrity of a dataset. It gives users the ability to edit or remove incorrect or malformed annotations. The user can also create new annotations to increase the size of the dataset.
VGG created this tool to be used with their VIA formatted object detection annotations. We modified the project to be compatible with Microsoft's Common Objects in Context(COCO) format.
Load Images
Import Annotation File
Edit Annotations
Save as COCO
VIA is an open source project released under the BSD-2 clause license.