Dimitris Tzionas
@dimtzionas.bsky.social
Assistant Professor for 3D Computer Vision at University of Amsterdam.
3D Human-centric Perception & Synthesis: bodies, hands, objects.
Past: MPI for Intelligent Systems, Univ. of Bonn, Aristotle Univ. of Thessaloniki
Website: https://dtzionas.com
3D Human-centric Perception & Synthesis: bodies, hands, objects.
Past: MPI for Intelligent Systems, Univ. of Bonn, Aristotle Univ. of Thessaloniki
Website: https://dtzionas.com
⚙️ CWGrasp:
👉 requires 500x less samples & runs 10x faster than SotA,
👉 produces grasps that are perceived as more realistic than SotA ~70% of the times,
👉 works well for objects placed at various "heights" from the floor,
👉 generates both right- & left-hand grasps.
🧵 8/10
👉 requires 500x less samples & runs 10x faster than SotA,
👉 produces grasps that are perceived as more realistic than SotA ~70% of the times,
👉 works well for objects placed at various "heights" from the floor,
👉 generates both right- & left-hand grasps.
🧵 8/10
March 14, 2025 at 6:36 PM
⚙️ CWGrasp:
👉 requires 500x less samples & runs 10x faster than SotA,
👉 produces grasps that are perceived as more realistic than SotA ~70% of the times,
👉 works well for objects placed at various "heights" from the floor,
👉 generates both right- & left-hand grasps.
🧵 8/10
👉 requires 500x less samples & runs 10x faster than SotA,
👉 produces grasps that are perceived as more realistic than SotA ~70% of the times,
👉 works well for objects placed at various "heights" from the floor,
👉 generates both right- & left-hand grasps.
🧵 8/10
👉 We condition both CGrasp & CReach on the same direction.
👉 This produces a hand-only guiding grasp & a reaching body that are already mutually compatible!
🎯 Thus, we need to conduct a *small* refinement *only* for the body so that its fingers match the guiding hand!
🧵 7/10
👉 This produces a hand-only guiding grasp & a reaching body that are already mutually compatible!
🎯 Thus, we need to conduct a *small* refinement *only* for the body so that its fingers match the guiding hand!
🧵 7/10
March 14, 2025 at 6:35 PM
👉 We condition both CGrasp & CReach on the same direction.
👉 This produces a hand-only guiding grasp & a reaching body that are already mutually compatible!
🎯 Thus, we need to conduct a *small* refinement *only* for the body so that its fingers match the guiding hand!
🧵 7/10
👉 This produces a hand-only guiding grasp & a reaching body that are already mutually compatible!
🎯 Thus, we need to conduct a *small* refinement *only* for the body so that its fingers match the guiding hand!
🧵 7/10
⚙️ CGrasp & CReach - generate a hand-only grasp & reaching body, respectively, with varied pose by sampling their latent space.
👉 Importantly, the palm & arm direction satisfy a desired (condition) 3D direction vector!
👉 This direction is sampled from ⚙️ ReachingField!
🧵 6/10
👉 Importantly, the palm & arm direction satisfy a desired (condition) 3D direction vector!
👉 This direction is sampled from ⚙️ ReachingField!
🧵 6/10
March 14, 2025 at 6:35 PM
⚙️ CGrasp & CReach - generate a hand-only grasp & reaching body, respectively, with varied pose by sampling their latent space.
👉 Importantly, the palm & arm direction satisfy a desired (condition) 3D direction vector!
👉 This direction is sampled from ⚙️ ReachingField!
🧵 6/10
👉 Importantly, the palm & arm direction satisfy a desired (condition) 3D direction vector!
👉 This direction is sampled from ⚙️ ReachingField!
🧵 6/10
⚙️ ReachingField - is a probabilistic 3D ray field encoding directions from which a body’s arm & hand likely reach an object without penetration.
👉 Objects near the ground are likely grasped from high above
👉 Objects high above the ground are likely grasped from below
🧵 5/10
👉 Objects near the ground are likely grasped from high above
👉 Objects high above the ground are likely grasped from below
🧵 5/10
March 14, 2025 at 6:35 PM
⚙️ ReachingField - is a probabilistic 3D ray field encoding directions from which a body’s arm & hand likely reach an object without penetration.
👉 Objects near the ground are likely grasped from high above
👉 Objects high above the ground are likely grasped from below
🧵 5/10
👉 Objects near the ground are likely grasped from high above
👉 Objects high above the ground are likely grasped from below
🧵 5/10
📢 We present CWGrasp, a framework for generating 3D Whole-body Grasps with Directional Controllability 🎉
Specifically:
👉 given a grasping object (shown in red color) placed on a receptacle (brown color)
👉 we aim to generate a body (gray color) that grasps the object.
🧵 1/10
Specifically:
👉 given a grasping object (shown in red color) placed on a receptacle (brown color)
👉 we aim to generate a body (gray color) that grasps the object.
🧵 1/10
March 14, 2025 at 6:35 PM
📢 We present CWGrasp, a framework for generating 3D Whole-body Grasps with Directional Controllability 🎉
Specifically:
👉 given a grasping object (shown in red color) placed on a receptacle (brown color)
👉 we aim to generate a body (gray color) that grasps the object.
🧵 1/10
Specifically:
👉 given a grasping object (shown in red color) placed on a receptacle (brown color)
👉 we aim to generate a body (gray color) that grasps the object.
🧵 1/10
If you are at #BMVC2024 , I will give a talk (remotely) at the ANIMA 2024 interdisciplinary workshop on "non-invasive human motion characterization".
👉 Room M2. Thursday at 14:00++ (UK).
👉 Talk title: "Towards 3D Perception and Synthesis of Humans in Interaction".
👉 Room M2. Thursday at 14:00++ (UK).
👉 Talk title: "Towards 3D Perception and Synthesis of Humans in Interaction".
November 28, 2024 at 2:01 AM
If you are at #BMVC2024 , I will give a talk (remotely) at the ANIMA 2024 interdisciplinary workshop on "non-invasive human motion characterization".
👉 Room M2. Thursday at 14:00++ (UK).
👉 Talk title: "Towards 3D Perception and Synthesis of Humans in Interaction".
👉 Room M2. Thursday at 14:00++ (UK).
👉 Talk title: "Towards 3D Perception and Synthesis of Humans in Interaction".