HO-3D is a dataset providing image sequences of various hand-object\ninteraction scenarios annotated with the 3D pose of the hand and the object and\nwas originally introduced as HO-3D_v2. The annotations were obtained\nautomatically using an optimization method, 'HOnnotate', introduced in the\noriginal paper. HO-3D_v3 provides more accurate annotations for both the hand\nand object poses thus resulting in better estimates of contact regions between\nthe hand and the object. In this report, we elaborate on the improvements to\nthe HOnnotate method and provide evaluations to compare the accuracy of\nHO-3D_v2 and HO-3D_v3. HO-3D_v3 results in 4mm higher accuracy compared to\nHO-3D_v2 for hand poses while exhibiting higher contact regions with the object\nsurface.\n