VisPose

VisPose is a tool for annotating 6D object poses and 2D bounding boxes in monocular image sequences. The tool provides an interface for projecting models into an image sequence and tweaking their fit before batch exporting COCO style annotations for the entire sequence. The tool supports fisheye sequences, but can only render visualizations of the bounding box annotations in the raw fisheye images.

Citation

If you use this dataset, we request you to cite the following work


@article{billings2020silhonet, title={SilhoNet-fisheye: Adaptation of a ROI based object pose estimation network to monocular fisheye images}, author={Billings, Gideon and Johnson-Roberson, Matthew}, journal={IEEE Robotics and Automation Letters}, volume={5}, number={3}, pages={4241--4248}, year={2020}, publisher={IEEE}

}

Examples

Fitting model into image sequence with fine-grained mouse control, using the VisPose annotation tool

Reviewing annotations with reviewer tool