Orientation with four discrete hypotheses and Gaussian mixtures
hardsubjective
Asked in interviews
An object's orientation is a state variable, but at each timestamp it can be one of 4 possibilities (each a 90° rotation apart) and you have no reliable noise information to pick one. How do you estimate the final orientation state, and how can Gaussian Mixture Models help?
Voice input needs Chrome or Edge. On this browser, type your answer — grading is the same.