Module P01009




UsingMatch Moving Technique to

Createa Character Moving From

OneObstacle to Another

Tableof Contents

1. Introduction 3

2. Aims and Objectives 6

2.1. Aims 6

2.2. Objectives 6

2.3. Learning Outcome 6

3. Methodology 8

3.1. What I am going to do 8

3.2. How I am going to do it 11

3.3. Justification. 13

4. Scope and constraints 15

5. Resources 16

6. Timetable 17

7. Reference list / Bibliography 19

1. Introduction

Thecomputer scientists have, since 1990s, made remarkable progress indeveloping algorithmic codes for autonomous agents. Such behaviorshave been instrumental in navigation in digital environments usingstrategies such as wandering, seeking, arriving, fleeing, evading oreven avoiding an obstacle. On the other hand, match moving has becomea major standard procedure for visual effects for all situationswhere CG and action materials get combined. This allows real andvirtual scenes that are composited together to enable them appearseamlessly as though from the same perspective[ CITATION Hor10 l 1033 ].

Thematch moving technique enables insertion of computer graphics to alive-action footage with scale, correct position, motion andorientation relative to the photographed objects within the shot[ CITATION Hir02 l 1033 ].Match moving have been primarily used in tracking camera movementsthrough a shot to enable an identical virtual camera to be moved inorder to be reproduced using the 3D animation program. Thecompositing of animated elements into original live action enhances aperfect appearance in a marched perspective hence appear asseamless. The benefits of the match moving are abundant providingcomprehensive expressive attributes to the motion pictures[ CITATION Hos01 l 1033 ].

Realisticcharacter motion is critical in media production like computer gamesand films. Use of lifelike characters produces special effects whichfacilitate the immersive experiences. Nevertheless, creation ofrealistic motions can be challenging. Matchmove is a mystery toseveral people[ CITATION Rey07 l 1033 ].This is because noticing a match move can be quite challenging. Matchmove takes the information from the real life setwhere a film is shot and recreate the camera while ensuring the focallength, tilt, height, motion and position relative to the subjectwithin the CG environment. Thereafter, after the creation of CGworld, the film is rendered or photographed using the virtual CG twinof real-life camera[ CITATION Sij05 l 1033 ].

Figure1:Lining up the virtual camera with the virtual set and the charactersduring the beginning of the shot.

Inthis way, CG elements created within the virtual world will possessthe same depth, perspective, same relationship and the same depth tothe moving camera similar to the live characters. This allowsseamless integration into the live plate ready for the final shot[ CITATION Wil14 l 1033 ].

Figure2: Creatingan animated CG camera matching all the properties of real-life camera

Aftercreation and approval of 3D camera solutions, there may appear otherelements to be matched. The interaction of the of full CG characterwith a live character requires addition of a virtual scene and themovements should be animated in order to match the movements of alive actor[ CITATION Men00 l 1033 ].The CG effects can also be applied to real life actor covered withreal life goo. In other cases, arms, heads or other body parts can bereplaced with the CG elements in order to seamlessly blend with liveaction plate[ CITATION Can12 l 1033 ].The animated CG puppets are required in casting the reflections andshadows on the CG elements added to the scene[ CITATION Par08 l 1033 ].

Gettingthe last 10-20 percent requires familiarity with match moving aspectssuch as technical camera aspects, data collection, and manipulationin virtual environments. The matchmover should demonstratecreativity of the highest order and know how to attack even withlittle or no information from the set. In addition, familiarity withenough cameras and lenses to enable spotting of any mistakes in dataset as well as account for[ CITATION Eva09 l 1033 ].

2. Aims and Objectives2.1. Aims

Thisresearch aims at using the match moving technique to create acharacter jumping from one obstacle to another

2.2. Objectives

  1. To animate an autonomous character within the dynamic environment

  2. To determine the effects of obstacle and character behavior

  3. To find out the feasibility of the jump with respect to animation parameters like length of the obstacle, locomotion speed and distance between the obstacle and the character.

2.3. Learning Outcome

Thisproject will enable me learn how to use match moving technique tocreate a character jumping from one obstacle to the other. Ianticipate that as this research will come to an end, I will be in aposition to produce realistic animation results by motion blendingusing warping techniques to identify and align the correspondencebetween motions. Also, I will be able to controlblending from the locomotion to jump using appropriate duration andtransition time. Through the generation of transition, I will be ableto further my skills in use of animation engines for both the jumpand locomotion. The survey in transition generation will reinforce mymotivation to formulate a methodology that automatically computestransition time and duration in ensuring coherence of the blendedmotions.

3. Methodology3.1. What I am going to do

Step1: Gather information

Forme to understand and decide on the research topic for thisdissertation, I was compelled to conduct a thorough research andevaluate various research topics and choose one from them. My passionfor match moving acted as the main driver to my choice of this topic.After the choice of the topic, I had to carry out the researchfurther to determine the expansive concept of match moving, what itentails and the necessary requirements. The research into matchmoving concept was reinforced by my familiarity with ‘StrangerThan Fiction’ which uses match moving extensively[ CITATION Hor07 l 1033 ].This implied that any project in the field of match moving is bothinformative and can be translated to profits when well mastered.

Step2: Explore

SinceI had the idea on match moving, I had to determine the match movingprocedure. I learned that the procedure entails

  1. Planning

  2. Gathering the data on the set

  3. Build the assets

  4. Review of data

  5. Attack the shot

Thisprovides the basic understanding of the match moving process.Majorly, I realized that tracking and calibration form the basicsteps in match moving[ CITATION Bon06 l 1033 ]. The identificationof the tracking features where the tracking algorithm locks onto andfollows using multiple frames (blips). Selection of the features ismainly because they have bright colored edges, corners and spotsbased on the algorithm used. Therefore, I wanted to focus on theselection of the features to create a character jumping from oneobstacle to the other. Thereafter, I had to determine the mostappropriate way to do the calibration in order to derive the motionof the camera using the 3D function. I realized that a camera is avector comprising elements of the position, focal length andorientation that is instrumental in defining the lighting on the filmplane[ CITATION Cho03 l 1033 ].

Step3: Work out any difficulties

Ican attest that that my topic was technical in a way and requiredthorough preparation. As a result, I had to identify any difficultieslikely to affect this study. This compelled me to researchextensively and determine the appropriate models, sketches, blueprints and approaches. From the literature reviewed, it was clearthat my research topic is s virgin area that is unfamiliar to manypeople. As a result, I had to gather ideas from different authors andcombine them in a way that could make sense and help me inunderstanding the match moving concept.

Step4: Determine my approach

Fromthe various approaches reviewed, the motivate system appeared to bethe most appropriate as it offers a novel way of generating contentusing the real-time motion synthesis and the dynamic event-basedprogramming. I learned that real-time motion synthesis presents a mixof motionwarping and kinematic methods. From the research, I discovered thatmotivate the approach allows for demonstrations through the use ofwalking gait variations like running, jumping and bounding[ CITATION Bro02 l 1033 ].This was critical since the study topic required creation ofcharacters jumping from one obstacle to the other.

Step5: Planning my scenes

Themain planning will be for motion and movements. In this case, I haveplanned to provide a path taking into account the avoidance problemof the obstacle. Thereafter, I will transform the path to thetrajectory that will return the locomotion parameters. I have decidedto use the probabilistic roadmap methods to help in random samplingof the configuration space and to construct the roadmap. This roadmapwill help in getting a collision-free 2D path in the 3D environment[ CITATION Guo96 l 1033 ].

Step6: Execution of Animation

AsI prepare for the project and determination of animation appropriateanimation, I plan to first give priority to the jump execution. Thiswill entail generation of the new jump with appropriate type andlength corresponding to locomotion parameter. Such an adaption iswell suited for game environment. After finishing the jump, I willgradually restore the locomotion parameters to their initial values.Also, the run-up duration will be determined to ensure the coherenceof the transition first based on constant speed variation takingconsideration of the induced acceleration. In this case, I havescheduled the transition to start after an event occurs during thelocomotion cycle.

Step7: Finishing Details and Refinements

Afterfinishing up with the animation, I will have to make some refinementsto the motion pictures and to record the results for the differentsituations as the character moves around the obstacle. In the finalrun-up, the speed necessary will entail very important accelerationthat will have to start from locomotion speed of the firstcharacter[ CITATION Par02 l 1033 ]. This will help evaluate thevariation in linear speed for the animation based on steering methodused.

3.2. How I am going to do it


Themethod to be used will be based on state-machine model comprising often states as presented in the figure below.

Figure3: Thestate machine for go-around or jump over obstacles

  1. The neutral state will correspond to freely moving environment for a character based on the motion captured data.

  2. The parameter vectors will allow for captured weights for the parameters of walking, running and speed.

  3. The jumping motions will be analogously generated through substitution of jump parameter length for speed.

  4. The other subparts will be composed of sub-parts responsible for generating transition from the locomotion to jump sequence

  5. The current locomotion will be adopted ensuring coherence with the requested jumps to provide variation for the locomotion in state 1.

  6. Blending will start after variation is finished at state 3

  7. State 4 will comprise of compatible support phase while state 5 will entail transition to walking or running

  8. At state 6, the locomotion will return to the neutral state through modification of motion parameters.

  9. States 7, 8, 9 and 10 will be specific to motion planning as soon as the obstacle appears in front of a moving character.

Figure4:Comparison of transition from running to walking and motion sequenceas the character gets across an obstacle

  1. The results will be dependent on current motion parameters, dimension of the obstacle and the distance from the character.

3.3. Justification.

Thechoice of the topic for creating of character jumping from oneobstacle to the other using match moving technique opted to uselocomotion and jumping engine since they enhance the production ofthe continuous motions that are parameterised. In this case, at anyrun-time, then the flat obstacles will be created and put in front ofa character. This will help in performing the transition from thelocomotion to jump using motion blending. I felt that this will beadvantageous when compared to the traditional approach that allow formanual entry of the duration. The variation of speed without the useof locomotion can be problematic. Hence, this approach will becrucial, especially when varying the take-off speed close to theobstacle.

Afterthe review of several research work related to my topic, I realizedthat most authors esteem the support phases in motion blending.Therefore, I felt that the choice of blending different motions likerunning and walking as the character goes from one obstacle to theother is necessary for determining the compatibility, and this wasnot general in the sense that the LHS support in walking motion wasincompatible with RTO for the jump. Therefore, the choice of thismethodology of locomotion and jump engine prevents such drawbacks byguaranteeing synchronous blending to determine duration and time oftransition.

4. Scope and constraints

  1. There is limited research which has been carried out on match-moving techniques in creating a character jumping from one obstacle to the other. This implies that I have to be tactful to get substantial information from the few sources available.

  2. The 10-state execution process chosen is simple to understand. Therefore, I can dedicate more time in final stages to enhance the quality

  3. The execution process for this project can be done using AutoDesk, hence the need to polish up on my Autodesk skills

  4. At first thought, it can be intuitively concluded that long jumps result in big errors. However, I believe that the error is not directly proportional the length of the jumps

  5. Use of the steering method may result in decreased speed as the character goes around an obstacle. This can be prevented by placing two-way points just near the obstacle in order to subdivide the complexity of the problem and help in recovering the original path.

  6. As I will be apprehensive of motivated model system, the quadratic variation will be in the position to enhance the flexibility and efficiency in handling the final and initial speeds that are identical.

  7. I think that the human eye may not be extremely sensitive to the acceleration discontinuities since it is anticipated that the results of this research will provide animations that are pleasant.

5. Resources

  1. Myself as an actor

  2. Lecturers and other relevant personnel as the reference

  3. Computers

  4. IT facilities and university library

  5. Online sources

  6. Secondary reference materials

  7. Printer


Academic calendar week no.


S1. 10

Start thinking about the topic and discuss with John as the relevant staff member.

S1. 11-12

Literature searching and model construction plans.

Christmas break

Review literature and start work on ideas.

S2. 2

Speak with the supervisor about topic and have it agreed upon.

S2. 3-8

Work on literature, ideas construction, and presentation plan/project plan.


Complete the project proposal plan.


Continue work on dissertation structure plan for chapters, do more on the models.


Complete final version – main chapters.

32rd May

Submit Interim Report.


Presentation/final submission.

7. Reference list / Bibliography

Bonsal, R. (2006). Stranger than Fiction? [AP-S Turnstile]. IEEE Antennas and Propagation Magazine 48(6), 143.

Brock, O., &amp Khatib, O. (2002). Elastic strips: a framework for motion generation in human environments. International Journal of Robots in human environments 21 (21), 1031–1052 .

Cantrell, B., &amp Natalie, B. (2012). Modeling the Environment: Techniques and Tools for the 3D Illustration of Dynamic Landscapes. Hoboken, NJ: Wiley.

Choi, M., Lee, J., &amp Shin, S. (2003). Planning biped locomotion using motion capture data and probabilistic roadmaps. New York, NY: ACM Transactions on Graphics (TOG).

Evans, D. J. (2009). Parameterized Computational Imaging: Optimized, Data Driven, and Time-varying Multiphysics Modeling for Image Extension. Moscow: University of Idaho.

Guo, S., &amp Roberge, J. (1996). A high-level control mechanism for human locomotion based on parametric frame space interpolation. Proceedings of Eurographics Workshopon Computer Animation and Simulation, pp. 97-113.

Hirabayashi, T. (2002). Study on the Control of the Variable-Structure-Type Locomotive Robot. 4th Report. Postual Change and Locomotion by Jumping of the Controlling Arm-Leg Type. Transactions of the Japan Society of Mechanical Engineers Series 58, 193-99.

Hornung, A., Ellen, D., &amp Leif, K. (2007). Character Animation from 2D Pictures and 3D Motion Data. ACM Transactions on Graphics 26(1), 1-6.

Hornung, E. (2010). The Art and Technique of Matchmoving: Solutions for the VFX Artist. Amsterdam: Focal.

Hoshino, J., Masanobu, Y., &amp Hirofumi, S. (2001). A Match Moving Technique for Merging CG Cloth and Human Movie Sequences. The Journal of Visualization and Computer Animation 12(1), 23-29.

Menache, A. (2000). Understanding Motion Capture for Computer Animation and Video Games. San Diego: Academic Press.

Park, J., &amp Bum-Jong, L. (2008). Vision-based Real-time Camera Match Moving Using a Known Marker . Optical Engineering 47(2), 27-31.

Park, W., Chaffin, D., &amp Martin, B. (2002). Modifying motions for avoiding obstacles. SAE Trans. 110(6), 2250-2256.

Reynolds, C. (2007). Steering behaviors forautonomous characters. Proceedings of Game Developers Conference Miller Freeman Game Group San Francisco, pp. 763–782.

Sijll, J. (2005). Cinematic Storytelling: The 100 Most Powerful Film Conventions Every Filmmaker Must Know. Studio City, CA: Michael Wiese Productions.

Wilk, S. (2014). Drawing for Landscape Architects . Berlin: DOM.