Polite Teacher
In construction site segmentation, like in many pixel-wise computer vision tasks, it is (relatively) easy to get the data. However, providing the labels is much more cumbersome. In this case not only because of the volume, but also due to specific construction expertise, which is required from the annotators.
As a consequence, one can have the “right” number of images for the training, but only a small fraction of them is labelled. Semi-supervised methods work exactly in such conditions.
We present Polite Teacher, a simple yet effective method for the task of semi-supervised instance segmentation. The proposed architecture relies on the mutual learning framework. To filter out noisy pseudo-labels, we use confidence thresholding for bounding boxes and mask scoring for masks. The approach has been tested with CenterMask, a single-stage anchor-free detector.
Tested on the COCO 2017 dataset, our architecture significantly outperformed the baseline at different supervision regimes. To the best of our knowledge, this was one of the first works tackling the problem of semi-supervised instance segmentation.
Bio
Adam Wiśniewski is co-founder and CTO of AI Clearing, based in Austin, Texas. AI Clearing is a leader in automating the supervision processes of large construction projects through AI computer vision tech. With a decade of hands-on experience in the construction industry, he has consistently delivered successful implementations of these cutting-edge technologies across multiple asset management projects for large-scale infrastructure initiatives. He co-founded and assumed the role of the CEO at Europe’s pioneering digital construction supervision platform. He was also associated for several years with PwC, delivering AI and reality capture solutions for heavy asset owners around the globe.