| |
| Figure 1 | Full body mesh topology (right : female body, left : male body). The gender-neutral face is zoomed for better visualization |
| Figure 2 | Levels of detail of Morgan - From left to right: High (50k), Medium (12k) and Small (5k) |
| Figure 3 | Front view of body semantic areas |
| Figure 4 | Back view of body semantic areas |
| Figure 5 | Front-view of the head. Top line: segmentation areas, bottom line: wireframe areas. (Left) complete head view with the face (white), left eye (red) and right eye (olive green) - (Middle) internal view without the face with the mouth bag (blue) - (Right) view without the face and the mouth bag - upper jaw (light green) and lower jaw (purple) |
| Figure 6 | Position of the 63 joints of Morgan |
| Table 1 | Morgan's Simplified and Simplistic skeleton joints nomenclature |
| Figure 7 | Morgan's face (right to left: front, back, and right profile) |
| Figure 8 | Morgan's face details ((a) mouth bag, (b) upper jaw, (c) lower jaw, (d) eyeballs) |
| Figure 9 | Morgan's eye details |
| Figure 10 | Morgan's eyes model |
| Figure 11 | Phase 1 timeline for MPEG avatar representation format |
| Table 2 | Requirements for Phase 1 of the MPEG Avatar Representation Format |
| Figure 12 | Avatar data model in ARF [25] |
| Figure 13 | Reference Avatar Pipeline in ARF |
| Table 3 | Facial Animation Sample format in ARF |
| Table 4 | Joint Animation Sample format in ARF |
| Figure 14 | 2D Avatar scenarios (top-left: Avatar call, top-right: Virtual anchor, bottom-left: Virtual customer service, bottom-right: Social media) |
| Figure 15 | 2D Avatar DNN model generation process (training) |
| Table 5 | Training input data for 2D Avatar DNN model generation |
| Figure 16 | 2D Avatar animation process (inference) |
| Table 6 | Possible input data combinations for 2D Avatar animation |
| Figure 17 | Avatar Reference Architecture |
| Figure 18 | Avatar workflows showing avatar functions performed by different entities |
| Figure 19 | Call flow for scene management in calls with avatars |
| Table 7 | Data Channel envelope format |
| Figure 20 | Mapping Avatar Functions to IMS DC Architecture |
| Figure 21 | Avatar generation call Flow |
| Figure 22 | Network Centric Call Setup and Capability Negotiation Flow |
| Figure 23 | UE Centric Call Setup and Capability Negotiation Flow |
| Figure 24 | IMS Avatar Delivery and Animation Flow |
| Table 8 | CRUD operations for Avatar management |
| Figure 25 | RTC Architecture for Avatar-based Communication |
| Figure 26 | Call flow for call setup and capability negotiation |
| Figure 27 | Call flow for Avatar Media and Metadata Exchange |
| Figure 28 | Generic Pipeline for Deep Learning-based 3D Mesh Watermarking |
| Figure 29 | Deep Learning-based 3D Mesh Watermarking for Avatar Protection |
| Figure 30 | potential avatar authentication procedure |