Ability to take any pretrained datasets, not just the "official" pretrained ImageNet weights.
- Train on SceneNet, results might even be better, as [1] suggests.
[1] McCormac, John, et al. "Scenenet rgb-d: Can 5m synthetic images beat generic imagenet pre-training on indoor segmentation." The IEEE International Conference on Computer Vision (ICCV). Vol. 1. 2017.
Ability to take any pretrained datasets, not just the "official" pretrained ImageNet weights.
[1] McCormac, John, et al. "Scenenet rgb-d: Can 5m synthetic images beat generic imagenet pre-training on indoor segmentation." The IEEE International Conference on Computer Vision (ICCV). Vol. 1. 2017.