Question about the MDA and DML

[zhang-qiang-github]

I have several questions about MDA and DML.

1. In DML, the key point is needed to calculate the loss. How to obtain the key point in the training? Does it comes from annotation information?
2. In MDA, how to use the fixed vertices? Is it used to split the contour to several segments, and use there segments to calculate loss? It seems an improvement for the Segment-wise Matching Scheme in DANSE . Am I correct?

[zhang-tao-whu]

For Q1:
The key point is obtained through the Douglas algorithm. Please refer to douglas.py.

For Q2:
Perhaps you can understand it this way. The difference between us and DANCE lies in the choice of alignment points. We use fixed points in a specific orientation, while DANCE uses the intersection points of contours and bounding boxes as alignment points. However, DANCE does not alleviate the unevenness in angle distribution.