2026-09-21T22:33:40Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/115430 2023-07-11 col_2142_5131 col_2142_10761 com_2142_5130 com_2142_10755 com_2142_234

Forsyth, David A 2022-05 application/pdf en text Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2022-11-11 without embargo terms The student, Victor Gonzalez, accepted the attached license on 2022-04-21 at 16:25. The student, Victor Gonzalez, submitted this Thesis for approval on 2022-04-21 at 16:31. This Thesis was approved for publication on 2022-04-25 at 15:25. DSpace SAF Submission Ingestion Package generated from Vireo submission #17892 on 2022-11-11 at 13:42:52 Improving 3D human pose estimation in-the-wild Gonzalez, Victor 2022-04-25 human pose estimation occlusion There has been some suspicion that 3D human pose estimation produces significantly worse results on in-the-wild images than on lab images. Confirming this suspicion is difficult, because it is hard to get 3D ground truth for in-the-wild images without measurement equipment significantly affecting the imagery. This thesis (a) demonstrates the suspicions are correct; (b) shows the effect is, at least in part, due to reconstructions not plausible (that is, "like" human poses); (c) explores simple augmentation can improve performance in situations with occlusion and (d) shows that natural methods to produce reconstructions that are plausible produce measurable improvements for in-the-wild reconstruction. Forcing methods to produce reconstructions that are plausible produces no major improvement on Human3.6M validation data; but this is because error on Human3.6M validation data is a poor predictor of error on in-the-wild data. This thesis shows that a registration error measure applied to reconstructions from multiple view data is a good predictor of ground truth error. Our registration error confirms that various procedures to enforce plausible reconstructions make notable improvements on in-the-wild error consistently across a number of distinct multiple view human action datasets. Thesis eng https://hdl.handle.net/2142/115430 Copyright 2022 Victor Gonzalez M.S. Thesis Computer Science University of Illinois at Urbana-Champaign Computer Science