Efficient 3D Implicit Head Avatar With Mesh-Anchored Hash Table Blendshapes

Figure 1. Visual Quality vs. Rendering Speed comparisons between our monocular 3D head avatars and prior state-of-the-art methods including NeRFBlendshape [14], INSTA [42], PointAvatar [41] and MonoAvatar [5]. Our approach achieves real-time rendering (i.e. > 30 mean FPS) with a 512 × 512 resolution, while produces comparable visual quality comparing to prior SOTAs [5], and gives significantly better results on challenging expressions than prior efficient avatars [14, 41, 42]. Webpage augmentedperception.github.io/monoavatar-plus