Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019]

Yuki Koyama
Yuki KoyamaPh.D. Student at The University of Tokyo
Precomputed Optimal One-Hop Motion Transition
for Responsive Character Animation
Yuki Koyama & Masataka Goto
The Visual Computer (CGI )
Target Problem
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
State Machine-Based Control for Interactive Apps
State machine-based control is widely used.
Each state is associated with a single motion clip (in the basic usage).
In response to run-time user input, the state is changed.
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Source (src) motion Destination (dst) motion
Transition
Triggered by user input

at an arbitrary frame
Goal: Achieve High-Quality Motion Transition
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Direct Approach (Current Practice)
■ Direct approach:
• Begin transition immediately
after receiving the trigger
• Directly interpolate the src and
dst motions
■ Problem:
• Visual artifacts (e.g., foot skate)
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Requirements for a New Method
■ A new method should
• Be compatible with the state
machine-based workflow
• Run in interactive real-time
apps
• Finish the transition as
quickly as possible
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Requirements for a New Method
■ A new method should
• Be compatible with the state
machine-based workflow
• Run in interactive real-time
apps
• Finish the transition as
quickly as possible
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Requirements for a New Method
■ A new method should
• Be compatible with the state
machine-based workflow
• Run in interactive real-time
apps
• Finish the transition as
quickly as possible
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Requirements for a New Method
■ A new method should
• Be compatible with the state
machine-based workflow
• Run in interactive real-time
apps
• Finish the transition as
quickly as possible
Quick
Overview of Our Method
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Our Approach: “One-Hop” Motion Transition
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Example of Our Results
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Overview of This Work
■ Technical Challenges:
. How should we choose a hop
motion and arrange it?
. How can we enable this in
real-time apps?
■ Our Solutions:
. Formulating a search
problem
. Solving the searches in
precomputation
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Overview of This Work
■ Technical Challenges:
. How should we choose a hop
motion and arrange it?
. How can we enable this in
real-time apps?
■ Our Solutions:
. Formulating a search
problem
. Solving the searches in
precomputation
Related Work
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Related Frameworks
[Holden+ , ][Kovar+ , ]
Statistical modelsMotion graphs
[Zadziuk , ]
Motion matching
➡ These are incompatible with the state machine-based workflow
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Related Frameworks
[Holden+ , ][Kovar+ , ]
Statistical modelsMotion graphs
[Zadziuk , ]
Motion matching
➡ These are incompatible with the state machine-based workflow
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Related Frameworks
[Holden+ , ][Kovar+ , ]
Statistical modelsMotion graphs
[Zadziuk , ]
Motion matching
➡ These are incompatible with the state machine-based workflow
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Related Frameworks
[Holden+ , ][Kovar+ , ]
Statistical modelsMotion graphs
[Zadziuk , ]
Motion matching
➡ These are incompatible with the state machine-based workflow
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Related Frameworks
[Holden+ , ][Kovar+ , ]
Statistical modelsMotion graphs
[Zadziuk , ]
Motion matching
➡ These are incompatible with the state machine-based workflow
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
State Machine-Compatible Approach [Ikemoto+07]
[Ikemoto+ ]
■ Similar to ours:
• They insert “hop” (intermediate)
motions to achieve natural
transitions
■ Key difference:
• We consider not only naturalness
but also quickness to finish the
overall transition (explained later)
1. Formulating a Search Problem
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
(Direct approach)
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
Unnatural
transition
(Direct approach)
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
hop
dst
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
More natural
transitions
hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
hop
Introducing

“stay durations”
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Even more natural
transitions
Trigger
src
dst
hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
One-Hop Transition Setting
Trigger
src
dst
hopOverall transition
duration
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Variables
Trigger
src
dst
hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Variables
Trigger
src
dst
hop
. Which motion clip to use as hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Variables
Trigger
src
dst
hop
. Which motion clip to use as hop
. Time offset of hop
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Variables
Trigger
src
dst
hop
. Which motion clip to use as hop
. Time offset of hop . Stay duration [src]
. Stay duration [hop]
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Objective
T⇤
t = arg min
Tt
{ S(Tt) + w R(Tt) }
<latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Objective
Smoothness cost:
Encourage each
transition to be as
natural-looking as
possible
T⇤
t = arg min
Tt
{ S(Tt) + w R(Tt) }
<latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">AAACPXicbZBLSwMxEMezPmt9VT16CRbBB5TdKigFoeDFg4eqrQrdumTTtAaz2SWZVcuyX8yL38GbNy8eFPHq1XTbQ30MhPwy/xky8/cjwTXY9rM1Nj4xOTWdm8nPzs0vLBaWls91GCvKGjQUobr0iWaCS9YADoJdRoqRwBfswr857OsXt0xpHso69CLWCkhX8g6nBEzKK9TrHlwlWyk+wC6we0iI6qbYDbj0EiMZTLBbcSv4bMM8Nwe8PbjusivD0xHVTb1C0S7ZWeC/4AyhiIZR8wpPbjukccAkUEG0bjp2BC0zC3AqWJp3Y80iQm9IlzUNShIw3Uqy7VO8bjJt3AmVORJwlh3tSEigdS/wTWVA4Fr/1vrJ/7RmDJ39VsJlFAOTdPBRJxYYQty3Ere5YhREzwChiptZMb0milAwhueNCc7vlf/Cebnk7JTKJ7vF6vHQjhxaRWtoAzloD1XREaqhBqLoAb2gN/RuPVqv1of1OSgds4Y9K+hHWF/fuYeqBQ==</latexit>
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Objective
Smoothness cost:
Encourage each
transition to be as
natural-looking as
possible
Responsiveness cost:
Encourage the overall
transition to finish as
quickly as possible
T⇤
t = arg min
Tt
{ S(Tt) + w R(Tt) }
<latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Objective
Smoothness cost:
Encourage each
transition to be as
natural-looking as
possible
Responsiveness cost:
Encourage the overall
transition to finish as
quickly as possible
Weight:
Control the
trade-off
T⇤
t = arg min
Tt
{ S(Tt) + w R(Tt) }
<latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Search Objective
Smoothness cost:
Encourage each
transition to be as
natural-looking as
possible
Responsiveness cost:
Encourage the overall
transition to finish as
quickly as possible
T⇤
t = arg min
Tt
{ S(Tt) + w R(Tt) }
<latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
Weight:
Control the
trade-off
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Evaluating Smoothness of Transitions [Kovar+02]
Dissimilar
➡ High cost for transition
(potentially non natural)
Similar
➡ Low cost for transition
(expected to be natural)
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Evaluating Smoothness of Transitions [Kovar+02]
High cost ➡ Unnatural transition
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Evaluating Smoothness of Transitions [Kovar+02]
Low cost ➡ Natural transition
2. Solving the Search Problem in Precomputation
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Precomputation
■ Solve the search problem for every possible trigger frame and
for every possible state transition
• May take a few days depending on the dataset size, but required
only once
■ Store optimal transitions in a look-up table
• In runtime, retrieve an optimal transition by using the triggered
frame as the key
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Precomputation
■ Solve the search problem for every possible trigger frame and
for every possible state transition
• May take a few days depending on the dataset size, but required
only once
■ Store optimal transitions in a look-up table
• In runtime, retrieve an optimal transition by using the triggered
frame as the key
Results
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Supposed Scenario: Interactive Dancing
Let a character dance using a game pad in an improvised manner
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Supposed Scenario: Interactive Dancing
■ Run in Unity
• It is straightforward to implement our method
using Unity’s PlayableAPI
■ Our dance motion dataset
• Around motion clips for hop candidates
• Each motion clip has around sec
■ Use frames (= . sec) for individual transitions
• I.e., src-to-hop ( . sec) and hop-to-dst ( . sec)
■ No postprocessing applied
• E.g., footskating removal (e.g., [Ikemoto+ ])
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Result: “Ghost Force” Disappears by a Standing-Up Motion
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Other Results
https://koyama.xyz/project/one-hop_motion_transition/
Discussions
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
The Direct Transition is Included in the Search Space
When the direct transition is already good, our solver is likely to select
the one-hop transition that is equivalent to the direct transition
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Future Work: Positional & Orientational Constraints
■ Current behavior
• Depending on the
inserted hop motion, the
position & orientation
may be changed
■ Future work
• Add constraints in the
search
Summary
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Summary
■ Optimal one-hop motion transition
• It inserts a hop motion between the
source and destination motions in
an optimal way
■ Consider both smoothness and
responsiveness
• Try to maximize both in balance
■ Can be precomputed
• Run in real-time apps
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Summary
■ Optimal one-hop motion transition
• It inserts a hop motion between the
source and destination motions in
an optimal way
■ Consider both smoothness and
responsiveness
• Try to maximize both in balance
■ Can be precomputed
• Run in real-time apps
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Summary
■ Optimal one-hop motion transition
• It inserts a hop motion between the
source and destination motions in
an optimal way
■ Consider both smoothness and
responsiveness
• Try to maximize both in balance
■ Can be precomputed
• Run in real-time apps
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Summary
■ Optimal one-hop motion transition
• It inserts a hop motion between the
source and destination motions in
an optimal way
■ Consider both smoothness and
responsiveness
• Try to maximize both in balance
■ Can be precomputed
• Run in real-time apps
The paper, videos, and slides are available at
https://koyama.xyz/project/one-hop_motion_transition/
Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019]
Appendix
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Changing Weight
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Future Work: Optimizing Blending Durations (c.f., [Wang+08])
Trigger
src
dst
hop
Blending durations

(we assumed fixed durations)
Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI )
Future Work: Blend Tree
Blend Tree:
Each state can also be
associated a dynamic
mixture of motion clips
Challenge: 

How can we
precompute all the
possible blends and
store the results?
Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019]
1 of 67

Recommended

[CHI 2018] OptiMo: Optimization-Guided Motion Editing for Keyframe Character ... by
[CHI 2018] OptiMo: Optimization-Guided Motion Editing for Keyframe Character ...[CHI 2018] OptiMo: Optimization-Guided Motion Editing for Keyframe Character ...
[CHI 2018] OptiMo: Optimization-Guided Motion Editing for Keyframe Character ...Yuki Koyama
326 views78 slides
Presentation1 by
Presentation1Presentation1
Presentation1Faye Mclellan
201 views8 slides
Web Animations by
Web AnimationsWeb Animations
Web Animationsbsuresh2
1K views45 slides
Datt 2501 week 11 by
Datt 2501 week 11Datt 2501 week 11
Datt 2501 week 11Joe Hambleton
562 views21 slides
Motion design in FIori by
Motion design in FIoriMotion design in FIori
Motion design in FIoriRoman Rommel
258 views23 slides
State of the Animation by
State of the AnimationState of the Animation
State of the AnimationRachel Nabors
6.7K views82 slides

More Related Content

Similar to Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019]

How To Make Multi-Robots Formation Control System by
How To Make Multi-Robots Formation Control SystemHow To Make Multi-Robots Formation Control System
How To Make Multi-Robots Formation Control SystemKeisuke Uto
8K views36 slides
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou... by
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...DeNA
945 views61 slides
GDC2011 - Implementation and Application of the Real-Time Helper-Joint System by
GDC2011 - Implementation and Application of the Real-Time Helper-Joint SystemGDC2011 - Implementation and Application of the Real-Time Helper-Joint System
GDC2011 - Implementation and Application of the Real-Time Helper-Joint SystemJubok Kim
2.5K views138 slides
A review of animation techniques for user interface design by
A review of animation techniques for user interface designA review of animation techniques for user interface design
A review of animation techniques for user interface designJean Vanderdonckt
1.5K views47 slides
Advanced animation techniques by
Advanced animation techniquesAdvanced animation techniques
Advanced animation techniquesCharles Flynt
3.8K views29 slides
Working with Amazon SageMaker Algorithms for Faster Model Training by
Working with Amazon SageMaker Algorithms for Faster Model TrainingWorking with Amazon SageMaker Algorithms for Faster Model Training
Working with Amazon SageMaker Algorithms for Faster Model TrainingAmazon Web Services
730 views44 slides

Similar to Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019](20)

How To Make Multi-Robots Formation Control System by Keisuke Uto
How To Make Multi-Robots Formation Control SystemHow To Make Multi-Robots Formation Control System
How To Make Multi-Robots Formation Control System
Keisuke Uto8K views
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou... by DeNA
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...
Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...
DeNA945 views
GDC2011 - Implementation and Application of the Real-Time Helper-Joint System by Jubok Kim
GDC2011 - Implementation and Application of the Real-Time Helper-Joint SystemGDC2011 - Implementation and Application of the Real-Time Helper-Joint System
GDC2011 - Implementation and Application of the Real-Time Helper-Joint System
Jubok Kim2.5K views
A review of animation techniques for user interface design by Jean Vanderdonckt
A review of animation techniques for user interface designA review of animation techniques for user interface design
A review of animation techniques for user interface design
Jean Vanderdonckt1.5K views
Advanced animation techniques by Charles Flynt
Advanced animation techniquesAdvanced animation techniques
Advanced animation techniques
Charles Flynt3.8K views
Working with Amazon SageMaker Algorithms for Faster Model Training by Amazon Web Services
Working with Amazon SageMaker Algorithms for Faster Model TrainingWorking with Amazon SageMaker Algorithms for Faster Model Training
Working with Amazon SageMaker Algorithms for Faster Model Training
User-friendly specification of robotic missions by Claudio Menghi
User-friendly specification of robotic missionsUser-friendly specification of robotic missions
User-friendly specification of robotic missions
Claudio Menghi33 views
Pokelabo flash 2013/4 by Shota Suzuki
Pokelabo flash 2013/4Pokelabo flash 2013/4
Pokelabo flash 2013/4
Shota Suzuki615 views
How I came to build a prototype for mobile develop by Rice Tseng
How I came to build a prototype for mobile developHow I came to build a prototype for mobile develop
How I came to build a prototype for mobile develop
Rice Tseng1.5K views
Virtual Simulation Of Systems by Hites
Virtual Simulation Of SystemsVirtual Simulation Of Systems
Virtual Simulation Of Systems
Hites998 views
The Web in Motion: animation's impact on UI and web design by Rachel Nabors
The Web in Motion: animation's impact on UI and web designThe Web in Motion: animation's impact on UI and web design
The Web in Motion: animation's impact on UI and web design
Rachel Nabors2.1K views
Day 1 presentation terminology by kelv_w
Day 1 presentation   terminologyDay 1 presentation   terminology
Day 1 presentation terminology
kelv_w862 views
Site Migration and Content Strategy by Randy Carey
Site Migration and Content StrategySite Migration and Content Strategy
Site Migration and Content Strategy
Randy Carey3.2K views
Core web vitals – Business impact and best practices - Meet Magento UK 2021 by Andrey Lipattsev
Core web vitals – Business impact and best practices - Meet Magento UK 2021Core web vitals – Business impact and best practices - Meet Magento UK 2021
Core web vitals – Business impact and best practices - Meet Magento UK 2021
Andrey Lipattsev119 views
Make your animations perform well - Anna Migas - Codemotion Rome 2017 by Codemotion
Make your animations perform well - Anna Migas - Codemotion Rome 2017Make your animations perform well - Anna Migas - Codemotion Rome 2017
Make your animations perform well - Anna Migas - Codemotion Rome 2017
Codemotion740 views

More from Yuki Koyama

[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G... by
[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...
[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...Yuki Koyama
90 views44 slides
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization by
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design OptimizationYuki Koyama
366 views90 slides
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics... by
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...Yuki Koyama
605 views70 slides
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati... by
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...Yuki Koyama
15.6K views91 slides
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン by
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザインYuki Koyama
316 views52 slides
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半 by
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半Yuki Koyama
1.9K views101 slides

More from Yuki Koyama(12)

[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G... by Yuki Koyama
[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...
[UIST 2022] BO as Assistant: Using Bayesian Optimization for Asynchronously G...
Yuki Koyama90 views
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization by Yuki Koyama
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
Yuki Koyama366 views
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics... by Yuki Koyama
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...
Decomposing Images into Layers with Advanced Color Blending [Pacific Graphics...
Yuki Koyama605 views
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati... by Yuki Koyama
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...
[SIGGRAPH 2017] Sequential Line Search for Efficient Visual Design Optimizati...
Yuki Koyama15.6K views
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン by Yuki Koyama
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン
[IPSJ-ONE 2017] コンピュテーションがつなぐ人とデザイン
Yuki Koyama316 views
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半 by Yuki Koyama
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半
[CEDEC2016] 大規模学習を用いたCGの最先端研究の紹介 - 前半
Yuki Koyama1.9K views
[CHI 2016] SelPh: Progressive Learning and Support of Manual Photo Color Enha... by Yuki Koyama
[CHI 2016] SelPh: Progressive Learning and Support of Manual Photo Color Enha...[CHI 2016] SelPh: Progressive Learning and Support of Manual Photo Color Enha...
[CHI 2016] SelPh: Progressive Learning and Support of Manual Photo Color Enha...
Yuki Koyama7.1K views
視覚デザイン探索のためのクラウドソーシングを活用したパラメタ空間解析(WISS 2014) by Yuki Koyama
視覚デザイン探索のためのクラウドソーシングを活用したパラメタ空間解析(WISS 2014)視覚デザイン探索のためのクラウドソーシングを活用したパラメタ空間解析(WISS 2014)
視覚デザイン探索のためのクラウドソーシングを活用したパラメタ空間解析(WISS 2014)
Yuki Koyama3.5K views
Crowd-Powered Parameter Analysis for Visual Design Exploration (UIST 2014) by Yuki Koyama
Crowd-Powered Parameter Analysis for Visual Design Exploration (UIST 2014)Crowd-Powered Parameter Analysis for Visual Design Exploration (UIST 2014)
Crowd-Powered Parameter Analysis for Visual Design Exploration (UIST 2014)
Yuki Koyama5K views
こだわり物理エンジン (Innovative Technologies 2013 @ DC EXPO) by Yuki Koyama
こだわり物理エンジン (Innovative Technologies 2013 @ DC EXPO)こだわり物理エンジン (Innovative Technologies 2013 @ DC EXPO)
こだわり物理エンジン (Innovative Technologies 2013 @ DC EXPO)
Yuki Koyama934 views
View-Dependent Control of Elastic Rod Simulation for 3D Character Animation (... by Yuki Koyama
View-Dependent Control of Elastic Rod Simulation for 3D Character Animation (...View-Dependent Control of Elastic Rod Simulation for 3D Character Animation (...
View-Dependent Control of Elastic Rod Simulation for 3D Character Animation (...
Yuki Koyama9.3K views
Real-Time Example-Based Elastic Deformation (SCA '12) by Yuki Koyama
Real-Time Example-Based Elastic Deformation (SCA '12)Real-Time Example-Based Elastic Deformation (SCA '12)
Real-Time Example-Based Elastic Deformation (SCA '12)
Yuki Koyama743 views

Recently uploaded

Piloting & Scaling Successfully With Microsoft Viva by
Piloting & Scaling Successfully With Microsoft VivaPiloting & Scaling Successfully With Microsoft Viva
Piloting & Scaling Successfully With Microsoft VivaRichard Harbridge
13 views160 slides
Five Things You SHOULD Know About Postman by
Five Things You SHOULD Know About PostmanFive Things You SHOULD Know About Postman
Five Things You SHOULD Know About PostmanPostman
38 views43 slides
"Node.js Development in 2024: trends and tools", Nikita Galkin by
"Node.js Development in 2024: trends and tools", Nikita Galkin "Node.js Development in 2024: trends and tools", Nikita Galkin
"Node.js Development in 2024: trends and tools", Nikita Galkin Fwdays
17 views38 slides
Ransomware is Knocking your Door_Final.pdf by
Ransomware is Knocking your Door_Final.pdfRansomware is Knocking your Door_Final.pdf
Ransomware is Knocking your Door_Final.pdfSecurity Bootcamp
66 views46 slides
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading... by
Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading...The Digital Insurer
24 views52 slides
Special_edition_innovator_2023.pdf by
Special_edition_innovator_2023.pdfSpecial_edition_innovator_2023.pdf
Special_edition_innovator_2023.pdfWillDavies22
18 views6 slides

Recently uploaded(20)

Piloting & Scaling Successfully With Microsoft Viva by Richard Harbridge
Piloting & Scaling Successfully With Microsoft VivaPiloting & Scaling Successfully With Microsoft Viva
Piloting & Scaling Successfully With Microsoft Viva
Five Things You SHOULD Know About Postman by Postman
Five Things You SHOULD Know About PostmanFive Things You SHOULD Know About Postman
Five Things You SHOULD Know About Postman
Postman38 views
"Node.js Development in 2024: trends and tools", Nikita Galkin by Fwdays
"Node.js Development in 2024: trends and tools", Nikita Galkin "Node.js Development in 2024: trends and tools", Nikita Galkin
"Node.js Development in 2024: trends and tools", Nikita Galkin
Fwdays17 views
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading... by The Digital Insurer
Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading...
Special_edition_innovator_2023.pdf by WillDavies22
Special_edition_innovator_2023.pdfSpecial_edition_innovator_2023.pdf
Special_edition_innovator_2023.pdf
WillDavies2218 views
ESPC 2023 - Protect and Govern your Sensitive Data with Microsoft Purview in ... by Jasper Oosterveld
ESPC 2023 - Protect and Govern your Sensitive Data with Microsoft Purview in ...ESPC 2023 - Protect and Govern your Sensitive Data with Microsoft Purview in ...
ESPC 2023 - Protect and Govern your Sensitive Data with Microsoft Purview in ...
Business Analyst Series 2023 - Week 3 Session 5 by DianaGray10
Business Analyst Series 2023 -  Week 3 Session 5Business Analyst Series 2023 -  Week 3 Session 5
Business Analyst Series 2023 - Week 3 Session 5
DianaGray10345 views
Case Study Copenhagen Energy and Business Central.pdf by Aitana
Case Study Copenhagen Energy and Business Central.pdfCase Study Copenhagen Energy and Business Central.pdf
Case Study Copenhagen Energy and Business Central.pdf
Aitana17 views
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院 by IttrainingIttraining
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院
TouchLog: Finger Micro Gesture Recognition Using Photo-Reflective Sensors by sugiuralab
TouchLog: Finger Micro Gesture Recognition  Using Photo-Reflective SensorsTouchLog: Finger Micro Gesture Recognition  Using Photo-Reflective Sensors
TouchLog: Finger Micro Gesture Recognition Using Photo-Reflective Sensors
sugiuralab23 views
STKI Israeli Market Study 2023 corrected forecast 2023_24 v3.pdf by Dr. Jimmy Schwarzkopf
STKI Israeli Market Study 2023   corrected forecast 2023_24 v3.pdfSTKI Israeli Market Study 2023   corrected forecast 2023_24 v3.pdf
STKI Israeli Market Study 2023 corrected forecast 2023_24 v3.pdf
The Forbidden VPN Secrets.pdf by Mariam Shaba
The Forbidden VPN Secrets.pdfThe Forbidden VPN Secrets.pdf
The Forbidden VPN Secrets.pdf
Mariam Shaba20 views
Data Integrity for Banking and Financial Services by Precisely
Data Integrity for Banking and Financial ServicesData Integrity for Banking and Financial Services
Data Integrity for Banking and Financial Services
Precisely29 views

Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation [CGI 2019]

  • 1. Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation Yuki Koyama & Masataka Goto The Visual Computer (CGI )
  • 3. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) State Machine-Based Control for Interactive Apps State machine-based control is widely used. Each state is associated with a single motion clip (in the basic usage). In response to run-time user input, the state is changed.
  • 4. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Source (src) motion Destination (dst) motion Transition Triggered by user input
 at an arbitrary frame Goal: Achieve High-Quality Motion Transition
  • 5. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Direct Approach (Current Practice) ■ Direct approach: • Begin transition immediately after receiving the trigger • Directly interpolate the src and dst motions ■ Problem: • Visual artifacts (e.g., foot skate)
  • 6. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Requirements for a New Method ■ A new method should • Be compatible with the state machine-based workflow • Run in interactive real-time apps • Finish the transition as quickly as possible
  • 7. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Requirements for a New Method ■ A new method should • Be compatible with the state machine-based workflow • Run in interactive real-time apps • Finish the transition as quickly as possible
  • 8. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Requirements for a New Method ■ A new method should • Be compatible with the state machine-based workflow • Run in interactive real-time apps • Finish the transition as quickly as possible
  • 9. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Requirements for a New Method ■ A new method should • Be compatible with the state machine-based workflow • Run in interactive real-time apps • Finish the transition as quickly as possible Quick
  • 10. Overview of Our Method
  • 11. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Our Approach: “One-Hop” Motion Transition
  • 12. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Example of Our Results
  • 13. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Overview of This Work ■ Technical Challenges: . How should we choose a hop motion and arrange it? . How can we enable this in real-time apps? ■ Our Solutions: . Formulating a search problem . Solving the searches in precomputation
  • 14. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Overview of This Work ■ Technical Challenges: . How should we choose a hop motion and arrange it? . How can we enable this in real-time apps? ■ Our Solutions: . Formulating a search problem . Solving the searches in precomputation
  • 16. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Related Frameworks [Holden+ , ][Kovar+ , ] Statistical modelsMotion graphs [Zadziuk , ] Motion matching ➡ These are incompatible with the state machine-based workflow
  • 17. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Related Frameworks [Holden+ , ][Kovar+ , ] Statistical modelsMotion graphs [Zadziuk , ] Motion matching ➡ These are incompatible with the state machine-based workflow
  • 18. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Related Frameworks [Holden+ , ][Kovar+ , ] Statistical modelsMotion graphs [Zadziuk , ] Motion matching ➡ These are incompatible with the state machine-based workflow
  • 19. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Related Frameworks [Holden+ , ][Kovar+ , ] Statistical modelsMotion graphs [Zadziuk , ] Motion matching ➡ These are incompatible with the state machine-based workflow
  • 20. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Related Frameworks [Holden+ , ][Kovar+ , ] Statistical modelsMotion graphs [Zadziuk , ] Motion matching ➡ These are incompatible with the state machine-based workflow
  • 21. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) State Machine-Compatible Approach [Ikemoto+07] [Ikemoto+ ] ■ Similar to ours: • They insert “hop” (intermediate) motions to achieve natural transitions ■ Key difference: • We consider not only naturalness but also quickness to finish the overall transition (explained later)
  • 22. 1. Formulating a Search Problem
  • 23. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst
  • 24. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst (Direct approach)
  • 25. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst Unnatural transition (Direct approach)
  • 26. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst
  • 27. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src hop dst
  • 28. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst More natural transitions hop
  • 29. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst hop
  • 30. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst hop Introducing
 “stay durations”
  • 31. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Even more natural transitions Trigger src dst hop
  • 32. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) One-Hop Transition Setting Trigger src dst hopOverall transition duration
  • 33. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Variables Trigger src dst hop
  • 34. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Variables Trigger src dst hop . Which motion clip to use as hop
  • 35. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Variables Trigger src dst hop . Which motion clip to use as hop . Time offset of hop
  • 36. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Variables Trigger src dst hop . Which motion clip to use as hop . Time offset of hop . Stay duration [src] . Stay duration [hop]
  • 37. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Objective T⇤ t = arg min Tt { S(Tt) + w R(Tt) } <latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
  • 38. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Objective Smoothness cost: Encourage each transition to be as natural-looking as possible T⇤ t = arg min Tt { S(Tt) + w R(Tt) } <latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
  • 39. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Objective Smoothness cost: Encourage each transition to be as natural-looking as possible Responsiveness cost: Encourage the overall transition to finish as quickly as possible T⇤ t = arg min Tt { S(Tt) + w R(Tt) } <latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
  • 40. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Objective Smoothness cost: Encourage each transition to be as natural-looking as possible Responsiveness cost: Encourage the overall transition to finish as quickly as possible Weight: Control the trade-off T⇤ t = arg min Tt { S(Tt) + w R(Tt) } <latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit>
  • 41. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Search Objective Smoothness cost: Encourage each transition to be as natural-looking as possible Responsiveness cost: Encourage the overall transition to finish as quickly as possible T⇤ t = arg min Tt { S(Tt) + w R(Tt) } <latexit sha1_base64="T4T3GkSkIxMtJugrGSYghaBpGts=">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</latexit> Weight: Control the trade-off
  • 42. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Evaluating Smoothness of Transitions [Kovar+02] Dissimilar ➡ High cost for transition (potentially non natural) Similar ➡ Low cost for transition (expected to be natural)
  • 43. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Evaluating Smoothness of Transitions [Kovar+02] High cost ➡ Unnatural transition
  • 44. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Evaluating Smoothness of Transitions [Kovar+02] Low cost ➡ Natural transition
  • 45. 2. Solving the Search Problem in Precomputation
  • 46. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Precomputation ■ Solve the search problem for every possible trigger frame and for every possible state transition • May take a few days depending on the dataset size, but required only once ■ Store optimal transitions in a look-up table • In runtime, retrieve an optimal transition by using the triggered frame as the key
  • 47. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Precomputation ■ Solve the search problem for every possible trigger frame and for every possible state transition • May take a few days depending on the dataset size, but required only once ■ Store optimal transitions in a look-up table • In runtime, retrieve an optimal transition by using the triggered frame as the key
  • 49. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Supposed Scenario: Interactive Dancing Let a character dance using a game pad in an improvised manner
  • 50. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Supposed Scenario: Interactive Dancing ■ Run in Unity • It is straightforward to implement our method using Unity’s PlayableAPI ■ Our dance motion dataset • Around motion clips for hop candidates • Each motion clip has around sec ■ Use frames (= . sec) for individual transitions • I.e., src-to-hop ( . sec) and hop-to-dst ( . sec) ■ No postprocessing applied • E.g., footskating removal (e.g., [Ikemoto+ ])
  • 51. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Result: “Ghost Force” Disappears by a Standing-Up Motion
  • 52. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Other Results https://koyama.xyz/project/one-hop_motion_transition/
  • 54. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) The Direct Transition is Included in the Search Space When the direct transition is already good, our solver is likely to select the one-hop transition that is equivalent to the direct transition
  • 55. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Future Work: Positional & Orientational Constraints ■ Current behavior • Depending on the inserted hop motion, the position & orientation may be changed ■ Future work • Add constraints in the search
  • 57. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Summary ■ Optimal one-hop motion transition • It inserts a hop motion between the source and destination motions in an optimal way ■ Consider both smoothness and responsiveness • Try to maximize both in balance ■ Can be precomputed • Run in real-time apps
  • 58. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Summary ■ Optimal one-hop motion transition • It inserts a hop motion between the source and destination motions in an optimal way ■ Consider both smoothness and responsiveness • Try to maximize both in balance ■ Can be precomputed • Run in real-time apps
  • 59. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Summary ■ Optimal one-hop motion transition • It inserts a hop motion between the source and destination motions in an optimal way ■ Consider both smoothness and responsiveness • Try to maximize both in balance ■ Can be precomputed • Run in real-time apps
  • 60. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Summary ■ Optimal one-hop motion transition • It inserts a hop motion between the source and destination motions in an optimal way ■ Consider both smoothness and responsiveness • Try to maximize both in balance ■ Can be precomputed • Run in real-time apps
  • 61. The paper, videos, and slides are available at https://koyama.xyz/project/one-hop_motion_transition/
  • 64. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Changing Weight
  • 65. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Future Work: Optimizing Blending Durations (c.f., [Wang+08]) Trigger src dst hop Blending durations
 (we assumed fixed durations)
  • 66. Yuki Koyama and Masataka Goto | Precomputed Optimal One-Hop Motion Transition for Responsive Character Animation | The Visual Computer (CGI ) Future Work: Blend Tree Blend Tree: Each state can also be associated a dynamic mixture of motion clips Challenge: 
 How can we precompute all the possible blends and store the results?