Scene Graph
A structured representation method in computer vision and robotics, in which graph nodes represent entities and edges represent relations between entities. It is invariant to pose.
Timeline (last 90 days)
The scene graph serves as a generative constraint that guides the sampling of diverse reset states, thereby covering object poses and grasp modes that do not appear in the video.
The scene graph provides dense rewards for each task stage, ensuring that when the policy starts from a reset state, the exploration path is sufficiently short and clearly guided.
A scene graph uses nodes to represent entities (such as objects, tools, and robot end-effectors) and edges to represent relations between entities (such as contact, above, and grasping), describing high-level semantic structure rather than dense numerical values.
Relational constraints often still hold when object positions change; for example, 'the cup is above the saucer' does not fail because the specific coordinates of the two change, which gives scene-graph-based policies a certain natural pose invariance.
A scene graph is a structured representation of image content in the form of a directed graph, with objects as nodes and inter-object relationships as edges.
All Facts (5)
The scene graph serves as a generative constraint that guides the sampling of diverse reset states, thereby covering object poses and grasp modes that do not appear in the video.
50%UnverifiedThe scene graph provides dense rewards for each task stage, ensuring that when the policy starts from a reset state, the exploration path is sufficiently short and clearly guided.
50%UnverifiedA scene graph uses nodes to represent entities (such as objects, tools, and robot end-effectors) and edges to represent relations between entities (such as contact, above, and grasping), describing high-level semantic structure rather than dense numerical values.
50%UnverifiedRelational constraints often still hold when object positions change; for example, 'the cup is above the saucer' does not fail because the specific coordinates of the two change, which gives scene-graph-based policies a certain natural pose invariance.
50%UnverifiedA scene graph is a structured representation of image content in the form of a directed graph, with objects as nodes and inter-object relationships as edges.
50%