Affordance-Guided Human-Object Interaction Motion Diffusion Diagram
Overview
A factorized two-factor diagram with three panels. It is an editable PowerPoint diagram adapted from the original paper figure under CC BY 4.0.
What the diagram shows
(A) Scene understanding: scene, planned object path and instruction feed an affordance denoising U-Net that outputs a task-specific affordance map, trained with ℒ_aff = ℒ_{x₀} + λ_BCE ℒ_BCE + λ_path ℒ_path. (B) Affordance interface: local voxel occupancy (P_pelvis, P_object, P_goal → V^pelvis, V^object, V^goal) and an affordance-voxel encoder. (C) Affordance-guided motion generation: a motion diffusion backbone with cross-attention (K, V from affordance features), test-time guidance, affordance-aware motion ranking with ℒ_rank, and an episode loop back to the scene. Every box, arrow, label and math term is a native shape you can recolor, resize and relabel for your own slides, posters or papers.
File format
The download contains .pptx (1 slide, 13.46 × 6.63 in; opens in PowerPoint, Keynote, WPS and LibreOffice Impress; Google Slides works too, but complex freeforms may shift slightly), .svg and .drawio. Includes preview.png (full resolution) and a 1200 px web preview.
Shapes and editing
The file has 104 native editable shapes: 27 auto-shapes, 37 text boxes, 4 freeforms, 36 lines/connectors. 15 embedded images (see below). You can edit all text and formulas, factor brackets, the three panels with title bars, U-Net / encoder trapezoids, P-labels, the MLP and Cross Attention blocks, dashed containers, arrows and the legend.
Embedded images and fonts
15 items are embedded images: the scene, voxel and motion renders. Text uses Times New Roman, which is preinstalled on Windows and macOS (LibreOffice uses the metric-compatible Liberation Serif). Sub- and superscripts are real formatted text runs, not equation objects. The SVG keeps text as text; the draw.io file is a shape-by-shape conversion and may need small nudges. This is a faithful redraw for reuse, not the authors' original source file.
Source & license Adapted from Figure 2 of MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances by Mingyuan Lei, Yoonchang Sung, Tat-Jen Cham (arXiv:2610.12416, https://arxiv.org/abs/2610.12416; code: https://github.com/LeiMingyuan/MAMHOI). Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Redrawn as editable shapes; changes made by 叫我Alonzo就好了/alonzolab. Not endorsed by the original authors. Attribution to the original paper is required when you publish or share the figure; the exact credit line is in LICENSE-ATTRIBUTION.txt inside each download.