Abstract This work strives for the classification and localization of human actions in videos, without need any labeled video training examples. Where existing relies on transferring global attribute or object information from seen to unseen action we seek classify spatio-temporally localize videos image-based only. We propose three spatial priors, which encode local person detectors along with...

The Transiting Exoplanet Survey Satellite (TESS) has an exceptionally large plate scale of 21"/px, causing most TESS light curves to record the blended multiple stars. This creates a danger misattributing variability observed by wrong source, which would invalidate any analysis. We develop method that can localize origin on sky better than one fifth pixel. Given measured frequencies (e.g., from...