- Developing and Evaluating Extensions to the 3D Cross Section of 4D Geometry
- Joe Subbiani
- 2377990s
- John Williamson
- 1 hour Create a unity project and do some research into ray marching.
- 2 hours Research into the derivation of signed distance functions (SDFs) for a series of shapes; a sphere, box, capsule, cylinder and torus.
- 1 hour Researching transformations that can be applied to SDF shapes. Includes translation, rotation, boolean operations.
- 1 hour Researching diffuse lighting for the ray marched scene.
- 0.5 hours Create a 3D scene containing all the shapes I had researched.
- 0.5 hours Use boolean intersection to find the 2D cross section of the 3D scene.
- 0.5 hours Implementing diffuse lighting into the 3D scene.
- 1 hour Created a working 4D sphere and a slider to travel along the W axis.
- 0.5 hours Created a working 4D box.
- 0.5 hours Implementing 2D rotation.
- *0.5 hours Successful experiment in implementing 4D rotation.
- 2 hours Researching and attempting to implement a 3D cone that would translate well to 4D. - Failure.
- 1 hour Researching and attempting to implement a 3D cone that would translate well to 4D. - Failure.
- 2 hours Researching existing SDFs for other shapes - namely an Octahedron and Tetrahedron.
- 2 hours Implementing a 4D octahedron and tetrahedron. - had to make some small tweaks but they work.
- 1 hour Implementing a 4D torus.
- 2 hour Setting up a scene to create a shape pose matching demo - it is really hard using the diffuse light. Would be better having each face coloured.
- 3 hours The shapes already used face normals and a coordinate system to colour each face, but the colour was dependant on the container, not the ray marched shapes orientation. Time was spent having the coordinate system rotate with the objects rotation. because of this I need a separate container to maintain 2 objects, one in the desired pose, and 1 for the user to manipulate.
- 2 hours Experimenting with a 3D "Shadow" of a 4D object, where the W axis is collapsed. Realised this is the same as taking the cross section at the center of the object.
- 1 hour writing up time log.
- 1 hour writing up status report.
- 1 hour created a new version of the 3D > 2D renderer using materials.
- 0.5 hours tried an onion skin effect, but will have to try a different way.
- 2 hour implemented materials for different objects.
- 0.5 hours Weekly meeting - direction to head in.
- 3 hour collecting research papers to look through to understand the field better
- read Arcball Rotation Control
- potential for 'commutative' rotation?
- read Rotating 3D shapes - Google Classroom
- Potential a simple and intuitive rotation solution
- read Arcball Rotation Control
- 1 hour Using Zotero to compile research papers.
- read Polyvision: 4D Space Manipulation through Multiple Projections.
- Motivation to have side by side slices of 4D space to see how the object will change.
- read Polyvision: 4D Space Manipulation through Multiple Projections.
- 1 hour Making a plan for tasks, representations and the shapes I intend to create for the user experiments.
- 2 hour
- read 3D Scene manipulation with 2D Devices and Constraints
- Less relevant than I thought it would be.
- Using real world constraints such as gravity and object intersection to manipulate objects in a 3D world with utilising all 6 degrees of freedom.
- read 4D Rendering: Projection & Cross Section
- Uses verticies, rather than ray marching (similar to Marc Ten Bosch).
- Seems easier to construct complex shapes.
- Harder to get a cross section.
- Cannot do curves well.
- Uses verticies, rather than ray marching (similar to Marc Ten Bosch).
- read 3D Scene manipulation with 2D Devices and Constraints
- 1 hour expanding the sources of relevant papers to find new material.
- 1 hour Read A visualization method of four-dimensional polytopes by oval display of parallel hyperplane slices
- very similar to my idea of a 4d cross section timeline and 4d layered cross sections.
- 0.5 hours Looked up shape name conventions.
- 0.5 hours organising Zotero with colour coding, to keep track of what I have read, need to read and favourites.
- 0.5 hours Weekly meeting - How to improve finding research papers - what to look for, and advice.
- take a look at psychology of understanding/teaching geometry.
- Reread relevant chapters in Things to make and do in the fourth dimension.
- may be good for explaining basic concepts?
-
2 hours Read multiple articles by Inigo Quilez
- good for getting to grips with ray marching.
-
0.5 hours Read The van Hiele Model of Geometric Thinking
- will be great dissertation material.
-
1 hours Read Keyboard-based control of four-dimensional rotations
- not super helpful, was hoping for something more interesting - rotates around 4th dimension
$\pi/16$ rad with key presses, not very precise.
- not super helpful, was hoping for something more interesting - rotates around 4th dimension
- 1 hour Writing C# script to control rotation of an object outside of the shader code.
- 3 hours Researching how to control the shapes without gimbal lock.
- 2 hours implemented 3D rotation by multiplying rotation matrices. axis don't mis-align, but gimbal lock is still a problem
- 2 hours attempted implemented 4D rotation by multiplying matrices. It was very messy and does not work very well...
- 0.5 hours weekly meeting - look into exp/log map rotation
- 3 hours attempted to implement a quaternion rotation system. wen't well until I tried it... going to need to work on that
- 2 hours implemented quaternion rotation. Very stable in 3D. Could not extend to 4D.
- 0.5 hours Reading about Marc ten Bosch's solution to 4D rotations
- 1.5 hours studying and digesting Geometric Algebra Rotors
- 2 hours studying to learn about rotors - Geometric Algebra
- 6 hours continuing studies on geometric algebra Geometric Algebra
- 2 hours begin implementation of a rotor class
- very slow process. I am struggling to translate mathematical concepts to code
-
1 hour make some developments on the rotor class
- I think I am heading in the right direction - each bivector - e.g
$B_xy$ is assigned the projected area of the parallelogram defined by the wedge product of 2 vectors. - the rotor produced from the geometric product
$ab = ab + a\wedge b$ can be built from the projected areas to define its orientation in space - still not sure how to produce the rotor
- not sure how to rotate about that rotor
- I think I am heading in the right direction - each bivector - e.g
- 4 hours Implement a quaternion based arc/grab ball to rotate shapes based on 3 circles
- works well, but need to implement a more fluid experience. The direction of rotation reverses if the object rotates 180deg in another direction.
- currently scrub left or right on screen to rotate whilst touching a circle with the mouse
- 3 hours research how you determine confidence/understanding in a learning environment
- definition of understanding
- visual aid helps rather than just words - i should do demos (already planned)
- 121 discussion is usually good, but maybe not feasible
- 3 hours working on rotors
- 0.5 hours weekly meeting
- discussion of how to implement rotors - need to implement a sense of direction
- 2 hours further attempts at implementing rotors - the shape does change, but very very weirdly
- 1.5 hours Creating a new unity project for final program
- Create a 3D and 2D shader side by side
- 2 hours Create a separate 2D shader and overlay it on the UI
- 2 hours Implement a script to control the cross section/object position
- 1 hours Creating a scalable UI
- 2 hours Create a 2 Layer transparency effect for the plane cross section
- 1 hour tidy up the UI and add a UI slider to control the object position
- 0.5 hours add other shapes and control them a slider
- 2 hours refining the 3D to 2D app and finishing it
- 0.5 hours weekly meeting
- discussion of Rotors, potentially use exponential and log addition of matricies for rotation
- linked to marc ten bosch Rotor3 implementation!
- 1.5 hours improvements to 2D to 3D demonstration
- 4 hours implementing 3D rotor using swipe gestures - still maintains axis during rotation so not intuitive
- for swipe rotation - each gesture will have to be a new rotor. I will have to create a new rotor each time, and add it to the rotor, without continuing rotation
- 1 hours implement 4D rotor (besides rotor * rotor product, and rotation of a vector by a rotor)
- 2 hours attempting to derive rotor * rotor product
- 2 hours trying to derive rotor * rotor product using the geometric product
- 3D rotation works, but cannot prove it works in 4D until 4D rotation function works - which i have not derived yet.
- 3 hours building a 4D to 3D demo, where the 3D object rotates to mimic the XW YW and ZW rotations.
- 2.5 hours implementing onion skin shader
- not as intuitive as I hoped for
- 1.5 hours implementing timeline shader
- x y z rotation - I can have the entire timeline rotate, or I can have each object rotate - not sure whats better
- 3 hours developing a menu system
- partially just to chill out, but also to learn more about UI and scene loading which will be important later for navigating the tests for users
- 3 hours completing the menu system
- not as refined as a final product but the menu navigation and carousel is working
- 1.5 hours transferring common code from shaders to an include file
.cginc- cannot transfer ray march functions unfortunately because GLSL does not use function pointers as it does not have a stack
- 1 hour implementing rotors to the grab ball
- Cannot transform rotors to unity
Transform.Rotate, so have to rotate the actual arc ball with quaternions
- Cannot transform rotors to unity
- 2 hours converting swipe rotation to use the global axis for intuitive rotation
- Realised I have to swap bxy and bxz or the rotation equation to use the local axes. This may have reprocusions for 4D
- 1 hour trying to have the grab ball rotation follow the mouse when tangent to the arc
- stuggling to work properly. restricting the axes do not seem to help. May need to base it off the quaternion
- 2 hours begin script plans for how to teach users
- 2.5 hours polishing grab ball to correctly use the tangent and to highlight the trajectory to move the mouse for users
- 2 hours create multi-rotation view of 4D (inspired by poly-vision), showing default view, xw, yw and zw. Used Cone and torus as example.
- also solved why cone was being unpredictable, but yet to implement in the main build.
- 3 hours Experiments with sdf polygon mesh to construct a pentachoron
- very glitchy in unity - triangles do not work
- 3 hours polish multi view representation and add it to the 4D shapes project
- 2.5 hours testing and implementing a pose matching script&scene that will randomly orient an object and check if 2 objects are of the same orientation
- 4 hours building a modular scene that will be used to conduct tests
- when specific options are set the scene can load each representation and the associated UI elements
- when specific options are set the scene can load the UI elements associated with each test
- 1 hour changing and improving the UI for "Shape Matching"
- 3 hours general improvements and polishing
- ensuring the UI is scalable to different sized screens
- ensure UI is clear and concise
- have not uploaded images yet
- 2.5 hours polishing modular scene
- reduced w axis timeline to only show 2 cross sections forward and backwards rather than 3 - totally 5 rather than 7
- have pose matching window be able to move anywhere on the screen, and is set to custom positions depending on the representation
- improved scale of UI elements
- 2 hours Add capsule to polished scenes, and 3D objects to
Shapes.cginc
- 0.5 hours weekly meeting
- 2 hours writing up intro script
- 3 hours re-organising plan for testing
- each stage of the test
- data to be collected
- 2 hours Creating some visualisations for the intro video for users
- 2 hours Trying to build a pentachoron
- 2 hours Created a working pentachoron (hyper tetrahedron)
- 2 hours playing with different "texturing"
- currently just colouring faces
- 3 hours deriving rotor-rotor product
- appears successful - interacts well with 3D rotation
- 2 hours created more visualisations for the users
- tesseract
- 2 hours derive 3D rotation expressions
- correct - but some signs are wrong, presumably because bosch takes Z as up, but the multiplication table, and me take Y as up
- 0.5 hours improving face colouring for "texturing"
- 3 hours finished deriving rotate function... it did not work
- 1 hour diagnose why it is not working, improved a bit but still not working as expected - I think something is wrong with the signs of each component of the equations
- 0.5 hours added another colour based texture - all positive red, all positive blue, overlayed over lighting
- 0.5 hours weekly meeting
- 1 hour 4D rotor works on 3D!, 4D rotation still doesn't work
- 1.5 hours re-derived 4D rotation - turns out I was doing P*uP rather than PuP*
- Rotation now works in yz, xz, xy AND 2 w3 planes. but not the last one... Need to re-derive Rotor3 rotation correctly and see if I learn anything.
- 0.5 hours considering how a grab ball should interact with XW, YW and ZW. Should it remain unmoving? or should I be controlling 2 3D grab balls to represent 3D and 4D
- 1.5 hours re-deriving rotor3 product and rotor4 rotation - found some stuff I thought maybe I missed but seems less accurate than before
- 1.5 hours working texture patterns
- Best Textures:
-
- Normal based diffuse lighting
-
- RGBW Lit
-
- Blue & Red normal direction with pattern
-
- RGBW with pattern
-
- Pentachoron is a bit funny, needs to have normal negated to show good results, shouldn't affect any other model whilst negative, just swipes direction of colours
- Best Textures:
- 1 hour further investigation into rotors
- found difference between local vs global rotation within rotor product. defined rotor product using "/" operator for global rotation - swiping
- 1.5 hours checking tri-planar rotation does not refer to planes of rotation - 4D texturing textures X Y Z and W not the associated planes.
- 1.5 hours more or less finished scripts
- 2.5 hours creating explanation animation for planar rotation.
- 0.5 hours Weekly meeting
- 2.5 hours re-deriving rotor equations using Johns multiplication table
- stably incorrect - I think I have had same behaviour before, but the equations seem nicer than my most recent attempts, even if it appears to be working less
- 1.5 hours implement toggle for shape options
- user can select which shape, and system can read selected shape
- 1 hour retry objects on UI with a second camera in the scene
- 4 hours tidy up the modular scene ALOT
- cameras for UI image displays
- redo all the UI positioning code
- set UI rescaling to only be called when the screen size changes - much less taxing
- create axis to show user what direction each of the 3 axes are pointing towards
- 1 hour implementing texturing in modular scene
- 4 hours implementing procedural file generation and json read/write
- each new file has a new id and contains the json data for the tests
- 3 hours Created a script that when triggered externally, will move the the next stage in the experiment.
- loop through representations
- loop through tests
- a lot of planning for how it will interact with other components i.e the object controller
- loop through representations
- 3 hours improvements to how data is handled and stored - each sub test stores the results in a new dictionary
- 0.5 hours Weekly Meeting
- 2 hours implementing new closed equation for rotors - THEY WORK (but are a bit glitchy)
- 2 hours reviewing rotor implementation to work out why it is glitchy. cannot find why. seems like it should work fine...
- 2 hours playing around with rotor to solve visual glitch - realised I was not normalising pseudo vector! IT WORKS - although sometimes shape scale up, but very rarely...
- 4 hours implementing swipe rotation in the modular scene
- several issues arose
- compiler could not replace rotation functions as there is so much content. had to reduce calls to rotate, but should be fine
- if the objects were shifted based on UI elements, textures would be offset and move across object
- I now have working swipe rotation
- several issues arose
- 0.5 hours converting generic manipulation script to use rotor, as the scenes to showcase each representation no longer work
- Not that this will be used in the experiment
- 1.5 hours create static class to contain game state between scenes.
- 1 hour moved swipe rotation code to add the ability to enable and disable swipe and grab ball rotation
- 1 hour implemented and improved grab ball rotation - was way over complicated before...
- 1 hours improve local rotation
- it was scaling weirdly, played around with it, found negating scalar and pseudo vector produces a much much more stable local rotation
- 1 hour make tetrahedron normals inside out to show pattern properly
- 1 hour creating and including icon to show center of rotation when using the swipe rotation along the Z axis
- 1 hour Recording audio for opening tutorial
- 4 hours editing and screen recording content to build up the intro tutorial video
- 3 hours realised local rotation still had issues and tried to solve them
- played around with it. I understand the problem but cannot yet come up with a solution
- may have to ditch the grab ball... Will keep trying though
- 0.5 hours realised rotation between 3D-2D and 4D cross section for YZ and XZ rotation was different in tutorial video...
- turns out if you base the rotation as a single rotation over
Time.time, the order of rotation actually matters even with rotors - so rather than a combined rotation it will do YZ THEN XZ. - by multiplying rotations over
Time.deltaTime, the order of rotations does not matter
- turns out if you base the rotation as a single rotation over
- 1.5 hours began plans to move TestController content to StateController - does not need to be checked every frame, much more robust, and can be checked between scenes.
- 0.5 hours rearranging UI controller into representation specific functions, with the plan to call them externally rather than every frame
- 0.5 hours Adding continuous rotation functionality to object controller
- 1 hour implementing test loading function in state controller
- still need to call UIController functions which are not static
- 1 hour experimenting with minimum angle between rotors
2 * Acos( (p * q.Reverse()).a )does not seem to work on normalised or un-normalised rotors...
- 0.5 hours moving the canvas scalar script into UI controller (simpler and more efficient)
- 0.5 hours adding initialisation of external non-static variables to static state controller class via another script.
- this will probably have to be moved around at a later date when I start actually loading and unloading scenes
- 2 hours Wrote up intro briefing for experiment using old main menu background. Added button to tutorial video - need to add start/pause buttons
- 0.5 hours rearranged project structure as a lot of directories with only a couple of files was making the project cluttered
- 2 hours building up tutorial video scene to have working scrub timeline and pause and play.
- You can now:
- read the experiment briefing, which takes you to
- tutorial video which you can pause and rewind, which takes you to
- modular scene to begin tests (this will be changed for an explanation scene)
- You can now:
- 0.5 hours test questionaire on paper - too many pages per participant
- 2 hours build survey scene where user can fill out answers between a test
- 1 hours add scripts for survey scene to read data
- 0.5 hour Create small script to initialise StateController on start up
- 1 hour Moving and modifying save functionality to state controller
- 4 hours
- Added button to go from survey scene to modular scene and save data
- Added button to go from modular scene to survey scene and save data
- Save selected shape and rotation data
- Solved issue with order of initialisation for StaticController
- Made 4D-3D counterpart also randomly rotated with main object
- 3 hours implementing progression from each test to each representation
- 0.5 hours implemented a better random rotation for rotation match
- 60% chance of 3D rotation
- 60% chance 1 of 4D rotation
- 100% chance of a 2nd 4D rotation
- 1.5 hours trying to find issue where the object continuously grows...
- xy and zw rotation together seems to cause the expansion, maybe I have copied something incorrectly - same behaviour has local rotation and part of the reason I needed to ditch it
- 0.5 hours implementing intro card when user is given a new representation
- 1 hour testing and improving
- 1.5 hours build project, fix issues with persistent data path + filename not combining properly
- 0.5 hours implemented timer
- if new representation it resets the timer when they finish reading the description of the representation
- 1 hour creating graph scene
- 1 hour improving graph scene to separate pose match
- 0.5 hours Weekly presentation
- 0.5 hours adding timer
- 1 hour running tests
- 1 hour polishing various components
- 0.5 hours weekly meeting
- 0.5 hours working angle between rotors
- 0.5 hours set up pop-up cards for each type of test
- 1.5 hours pause timer when pop up card is visible and activate/disable cards explaining the test at the beginning of each test
- 1 hour add time limit for each test, and set a different number of iterations for each type of test
- 1.5 hours implement graph functionality based on JSON data
- 3 hours experimenting with pattern texture on all elements of multi-view
- very chaotic, but does provide extra information
- 0.5 hours creating new logo
- 1.5 hours trying to find problem with rotors AND FIXING THEM
- Rotors seem to 100% work now!
- 0.5 hours Added torus with spherical crossSection to
- 1 hour Navigation from Survey to Graph scene when new representation
- 2.5 hours playing with patterns on multi-view and timeline (multi-view was actually wrong)
- 3 hours playing around with patterns on multi-view and timeline trying to fit the patterns in the right place
- super slow as compiling the complex shaders takes ages
- 2.5 hours finally finished getting pattern on timeline and multi-view correctly
- was using W separation for ages rather than X separation between objects
- 1 hour debugging why it was not saving correct file
- 3.5 hours Implemented Practice Scene
- 0.5 hours Added Play Icon on play button for video
- 2.5 hours Fixing bug where practice was skipped
- 1.5 hours Running practice with Sophie and making improvements
- reword some of intro
- not enough focus on object cross section in video
- space bar not explained in practice scene
- space bar removing selected option through automatic navigation
- not enough time in practice (add 30 secs)
- hard to understand 4D Multi-view description - add example
- improve wording in rotation_match intro
- explain within plane 2D rotation
- better 2x 4D planes of rotation explanation
- reword pose_match survey
- disable movement in rotation_match
- 0.5 hours Show graph at end, before thank you
- 2 hours implementing saving results via email as WebGL does not have a file system, it is temporary
- 1 hour testing out itch.io - video player doesn't work...
- 0.5 hours weekly meeting
- 0.5 hours polishing - adding copy to clipboard and improving intro scene
- 1 hour writing up rotor in python and creating data analysis jupyter notebook
- 1 hours writing up data analysis tool
- 1.5 hour Adding data to be saved and creating a variant of the app without the tutorial video
- 0.5 hours setting up calender booking with calendly
- 1 hour testing WebGL without video
- works great except...
- cannot copy to clipboard
- cannot email
- Overlapping UI Element in Graph scene - fixed
- 1 hour alternative WebGL specific emailing
- still does not work with itch.io
- 1.5 hours researching other places to host game
- new grounds - email works
- 0.5 hours transferring questionnaire from google docs to google forms and creating a checklist for when running the experiment
- 0.5 hours rewriting script for tutorial covering rotation
- 1 hour implementing working copy and paste input box using someone elses plugin
- 1.5 hours testing copy and paste works
- copy and paste doesn't work in full screen for some reason
- email still doesn't work on eduroam - but i have copy/paste backup
- 2 hours recording improvements to tutorial video
- 1 hour editing
- 1 hour handwriting captions
- 2 hours reading other dissertations and structure my dissertation
- created a separate repo for final stuff + dissertation
- 2 hours adding more to dissertations
- 1 hour writing up weekly meeting slides mostly for noting down questions about dissertation
- 2.5 hours begin presentation outline
- 1 hour Test system twice and save data
- 1.5 hours Adding gifs to presentation
- 0.5 hours Change lock rotation keys around for x and y in 3D
- 1 hour port Google Slides presentation to power point to have an actual source file
- changed theme and moved things around with it
- 0.5 hours redo lock rotation keys as it was wrong before and I hadn't uploaded it to newgrounds
- 3 hours writing up shape match data analysis functions
- 0.5 hours Weekly meeting
- 1 hour adding additional data to experiment
- count use of w and use of swipe
- 1.5 hours Experiment 1
- went smoothly, on schedule
- worked on data analysis too
- 1 hour Data analysis
- generic functions to understand data
- 4 hours Experiment 2 and 3
- went smoothly, on schedule
- worked on data analysis too
- 5 hours Data analysis
- more generic functions plus graphs - box and whisker plots for correctness/accuracy and time
- 2 hours Experiment 4
- seemingly more laggy from the sounds of it
- 4 hours Data Analysis
- covariance matrix / correlation coefficient matrices to interpret shape data
- 1.5 hour Extending correlation coefficient matrices for shape match data and copy the same process to rotation match data
- 0.5 hour Splitting up separate data analysis files
- single file
- multiple file
- rotor derivation
- 2.5 hours building multi-file analysis by compiling each file into one well formatted file
- 0.5 hours weekly meeting
- 2 hours Experiment 5
- 3 hours Data analysis on multiple files for shape match
- all but correlation
- 1 hour Writing up intro to dissertation and adding citations for stuff to include
- 0.5 hours Adding to data analysis
- 0.5 hours making weekly presentation slides
- 2 hours finish implementing non-correlation based data for multiple files
- also created an averaging system for difference between rotors
- 0.5 hours weekly meeting
- 5 hours shape matching correlation matrices
- had to redo as accidently deleted
- 2 hours shape match correlation graphs
- 1 hour Experiment - technical difficulties will reschedule
- 1.5 hours implement correlation graphs for rotation matching
- 1 hour implement correlation for pose matching
- 1 hour improvements to data analysis
- 2 hours Experiment 6
- 2 hours Discussion of problems with rotation in the dissertation
- 2.5 hours Explanation of geometric algebra in dissertation
- 1.5 hours discussing the different approaches to rendering 3d objects
- 2 hours Experiment 7
- 2 hours Experiment 8
- 1 hour Rsearching stereographic projection
- 0.5 hours weekly meeting
- 1.5 hours discuss achievements and shortcomings of previous works
- 3 hours dissertation ray marching
- 2.5 hours creating ray marching figures with blender and illustrator
- 1 hour write about mesh vs ray march cross section
- 0.5 hours adding questionnaire pdf
- 0.5 hours weekly meeting
- 1.5 hours experiment 9
- 1 hour improve aims in dissertation
- 1 hour discuss drawbacks and qualities of polyvision, and other improvements
- 0.5 hours improve ray marching diagram with labels
- 1 hours summarise 5 levels of geometric understanding
- 1 hour explain moscow
- 0.5 hours correct mistake on ray marching diagram
- 1 hour read some papers on 3D interaction
- 1 hour improvements around MoSCoW
- 1.5 hours summarise limitations of project and must haves
- 1.5 hours create figures for timeline and multi-view for dissertation
- 1 hour add labels to figures and add them to the dissertation
- 1.5 hours finish analysis and requirements chapter
- 1 hour add draft rotor derivations to implementation section
- 0.5 hours data analysis stacked bar charts for shape match
- 0.5 hours initial rotation distance from 90 degrees
- 1 hour preliminary evaluation for shape match
- 2 hours preliminary evaluation for rotation and pose matching
- 0.5 hours Summary for preliminary evaluation
- 1 hour tidying rotor equations in dissertation
- 0.5 hours weekly meeting presentation
- 0.5 hours weekly meeting
- 1 hour making tense consistent
- 0.5 hours make correlation red to blue
- 1.5 hours experiment 10
- 0.5 hours reviewing more of dissertation for wording and correct tense
- 1.5 hours writing up the 5 stages of the learning process
- 1.5 hours complete section on van hiele model
- 0.5 hours revise ordering of aims
- 0.5 hours move rotor equations to appendix
- 1.5 hours add graphs and rotors to presentation slides
- 0.5 hours use clever references
- 2 hours describe SDF shadows and lighting
- 0.5 hours add custom data types to listing syntax highlighting
- 4 hours derivation of 4D shapes SDFs
- 3 hours Explain projection and components of multi-vectors
- 2 hours explain derivation of rotor equations
- 2 hours create figures for projections
- 6 hours explain timeline, multi-view and onion skin + add figures
- 1.5 hours finish talking about 4D-3D
- 0.5 hours weekly meeting
- 0.5 hours restructuring
- 1 hour figures for 4D-3D
- 2 hours experiment 11
- 1 hour add figures for 4D - 3D
- 0.5 hours begin design chapter
- 2 hours discussion of tutorial video
- 2.5 hours* discuss experimental design
- 2.5 hours discuss design of manipulation
- 3 hours texturing of objects section
- 2 hours data collection section
- 0.5 hours weekly meeting
- 3.5 hours improvements to implementation section
- design json data
- compact language
- explain choice of 4D primitives
- explanation of components of project
- 1.5 hours Begin presentation script
- 2 hours writing readme.md and manual.md + improvement to Rotor python module
- 2 hours common approaches to 3D manipulation
- 3.5 hours discuss rockin' mouse, virtual sphere and begin discussing arc ball
- 3 hours finish methods of rotation and therefore finish backgrounds chapter
- 2 hours write up preliminary experiments
- 1 hours begin intro to rotation mechanics
- 4 hours trim down and tidy background chapter
- 1.5 hours tidy introduction chapter
- 1.5 hours tidy requirements chapter
- 3 hours write up tasks and parameters section
- 0.5 hour figures for experiment tasks
- 2.5 hours explain swipe rotation implementation
- 3 hours explain grab ball rotation implementation
- 3 hours further experiment with grab ball
- test implementation with interactable grab ball
- test SDF grab ball
- 0.5 hours grab ball figures
- 1 hour finish grab ball section, finished implementation chapter
- 2.5 hours tidy background chapter
- 2 hours write up skeleton for evaluation
- 0.5 hours reword some stuff in introduction
- 2.5 hours write up limitations section
- 0.5 hours Weekly meeting
- 4 hours record and edit presentation animations
- 1 hour add presentation to LFS
- 2 hours working on presentation script
- 2 hours Add and improving wording of presentation
- 0.5 hours Improve hyper links and figures in dissertation
- 3 hours Discuss shape matching results and jaccard index
- 1 hour improvements to background chapter
- 2 hours discuss representations effectiveness for rotation match
- 3.5 hours finish discussing rotation match results
- 2 hours polish results section and arrange figures
- 3.5 hours finish pose matching results section
- 1 hour 4D to 3D abstraction representation, qualitative data
- 2 hours complete conclusion
- 3 hours improve grammar throughout
- 2.5 hours proof read implementation chapter
- 3 hours proof read and tidy evaluation and conclusion chapters
- 3.5 hours finish dissertation
- finish analysis and discussion
- finish abstract
- 2 hours finish slides for presentation
- still need to write notes and record
- 0.5 hours weekly meeting
- 3 hours make suggested improvements
- 2 hour finish script for presentation
- 1.5 hours record presentation
- 3 hours editing presentation
- 3 hours editing presentation - finished
- 0.5 hours making final improvements to the dissertation