# Augmented Reality ## 01 - Introduction ### Mixed Reality 1. **What is Mixed Reality?** - Presenting both, real world and virtual world objects in a single display - Umbrella term for: - Augmented Reality - Augmented Virtuality - Virtual Reality 2. **What is Augmented Reality?** - virtual enhancement of the real world - user sees real world - with virtual objects placed in real world 3. **What is Augmented Virtuality?** - real enhancement of a virtual world - real data is mapped on virtual objects (textures etc.) - >50% of the world is computer generated 4. **What is Virtual Reality?** - fully virtual world - using physical laws/world in virtual environment - intent is to immerse the user 5. **Describe Milgrams Reality-Virtuality Continuum** - Range from real environment to a purely virtual environment - everything in between is mixed reality 6. **What are the AR Characteristics?** - AR enhances users perception and interaction of the real world - uses: - head mounted displays (HMD) - head-up displays (HUD) - smartphones - not limited to display technologies - integrates additional info in real world - tangible interfaces - intuitive interaction with real life objects - mobile or portable 7. **Tell differences between AR and VR** - Tracking: AR high accuracy/VR low okay - Display Device: AR small view/VR wide field of view - Scene Generation: AR minimal rendering ok/VR realistic images required ### Classification 8. **Describe the term EWK** - Extent of World Knowledge - Knowledge of the displayed world - where, what, where and what 9. **Describe the difference between an unmodelled and modeled world** - unmodelled world could contain images or objects - modeled world contains info about object positions, viewpoint observation, user interaction 9. **What does the term RF mean?** - Reproduction Fidelity - describes the quality in which the display can reproduce the content - simple wireframes to hi-fidelity 3D animations - video to high res 3D 10. **What does the term EPM mean?** - Extent of Presence Metaphor - how much the user is intended to feel present - different ways to display the environment - Monitor to HMD 11. **Explain the Milgram-Weiser Continuum** - Combination of Milgram and Weiser Continuum - Ubicomp vs Monolithic - Real vs Virtual Environment - Mobile Gaming more virtual and ubicomp ### Structure of MR Systems 12. **Name the main issues of AR** - Calibration - Registration - Synchronisation - Tracking - Visual coherence - Real-time - mobile requirements - Processor power limited 13. **Name the main issues of VR** - Latency - Lag - Degree and metaphor of interaction - Navigation - Cybersickness 13. **Name the components of the sensor feedback loop** - **User** gives - **Input** to the - **Reality Engine**. The Engine makes changes and - displays the results on a **display** ### Categorising Mixed Reality 14. **Name the different Categories of MR** - Telepresence - AR - Desktop AR/VR - Fishtank VR - Immersive VR 15. **Explain Telepresence** - display of the real world - audio and video data - places user in another real existing place - eg. robot control 16. **Explain AR** - real world enhancement with additional content - HMD, smartphones,... - position tracking required (and often an issue!) 17. **Explain Desktop AR/VR** - PC based - 3d-glasses in front of display - cheap 18. **Explain Fishtank VR** - Projection screens - display on small display area and for more users 19. **Explain Immersive VR** - Goal: Immersion - CAVE/HMD/curved screen - user as far as possible isolated from the real world - multiuser possible - = VR - full virtual world - Cybersickness! ### Different types of AR 20. **Name the different types of AR** - Head-up displays (automotive) - screen based (smartphone) - SAR = Spatial Augmented Reality (Textile, Automotive) - stationary (marketing, lego) - HMD-based (training, maintenance, assembly lines) - smart glasses (logistic, military) - smart lenses (replace HMDs) 21. **Name examples for every field of AR in real life** - archaeology: documentation, cultural heritage - Medicine: training, visualisation - Entertainment: Smartphones, consoles - tourism: info visualisation - marketing: ads, catalogues - logistics: guidance systems - manufacturing: manuals, instructions ### Challenges and Development 22. **Name most common problems and challenges** - graphics rendering HW and SW - Tracking techniques (changes of viewer position properly reflected by rendered graphics) - Tracker calibration and registration tools (precision) - display hardware (virtual to real world) - computer processing hardware for AR - interaction techniques 23. **Name current focuses in development** - input devices and displays - improving optical tracking technology - research and evaluation of displays - software --- ## 02 - Computer Graphics ### Introduction 24. **What are computer graphics?** - graphics created using computers, the representation and manipulation of pictorial data - opposite of computer vision ### Color Perception 25. **What is responsible for Color Perception?** - different cones (RGB) ### Color Models 26. **Name the different Color Models** - RGB - CIE - HSV - HLS - CMYK ### Graphics Pipeline 27. **What is the goal of the Graphics Pipeline?** - displaying 2D projection in 3D scene - subsequential stages transforming data frame stage to stage - geometric and rendering processes involved 27. **What are the steps of the Graphics Pipeline?** - Vertices to 3D geometric primitives - Modelling and transformation - Lightning - Clipping - Rasterization - Texturing and shading 28. **Name the basic primitive objects** - points - lines - triangles 29. **Out of what are all 3D objects constructed?** - mesh of triangles 29. **What do triangles typically have?** - 3 Vertices interconnected by edges - 1 normal vector indicating the front side - Additionally normal vectors for vertices exist 28. **Name the different 3D primitives** - Sphere - box - cylinders - torus (Donut shaped form) 29. **Explain what happens in the modeling and transformation stage** - primitives are transformed from the local coordinate system to the world coordinate system 30. **What is a transformation?** - Scaling - rotation - translation - as a 4x4 matrix 31. **What coordinate systems exist for 2d and 3d?** - 2D - Screen Coordinate System - Viewport Coordinate System - 3D - World Coordinate System - Object Coordinate System - Left vs Right handed system 31. **What is a Frustum and what is Frustum Culling?** - Is a volume where its contents are projected for rendering - Has a near and a far clipping plane - Culling means everything outside the area of the frustum is not rendered 31. **How do changes of FOV and aspect ratio affect an image?** - High vertical/horizontal FOV make things smaller - Different aspect ratios enlargen things on the corresponding side 32. **Which different type of lightning models exist?** - Global lightning - Local lightning (indirect) 33. **What different type of light sources exist?** - directional light - spot light - point light ### Shading 34. **Name the different types of shading and explain it in short terms** - Flat shading: norm-vector + color value (once per frame) - Gouraud shading (intensity interpolation shading): Vertex + color interpolation - Phong shading: Vertex normal vectors + color interpolation per pixel basis ### Texturing 35. **How can we change the surface of a 3D model?** - Applying images on the 3D models/objects - coordinates of texture are mapped onto the object ### Z-Buffer 36. **What is a Z-Buffer?** - = stores distance from image plane on a per pixel basis - expands between near and far clipping plane - nonlinear - higher precision for pixels closer to near plane 37. **Another name for Z-Buffer** - Depth buffer 37. **What is Z-Buffer fighting?** - Z-Buffer fighting is caused by close by polygons - Usually results in flickering of the two objects where the intersection is ### Rasterization 38. **What happens in the rasterization step?** - = scan conversion - 3D model of the scene is rasterized and a 2D image is generated - Colors from shading are evaluated - Textures applied - surface in front based on Z-Buffer are drawn ### High Dynamic Range 39. **What is HDR?** - mapping the full color range - for displaying on a screen not capable of showing the full color range - = Tone mapping ### Basic Optimization 42. **Name different types of optimization operations** - Anti-aliasing (removing sampling artifacts - level of detail (less detail, better performance), can be done beforehand as dynamical processing is intense and error prone - MIP Mapping (level of detail applied on textures) - Billboards (simple representations of complex geometric objects) ### Scene Graphs 42. **What is a scene graph?** - Collection of nodes in a graph or tree structure - Node may have many children but only a single parent - effect/changes of a parent is applied to all its child nodes - Leaf nodes contain geometry ### Culling Techniques 42. **What are the most common culling techniques?** - Frustum culling - Backface culling (typ. biggest performance increase) - Occlusion culling - Contribution culling 43. **How can Frustum Culling be improved?** - By using bounding volumes instead of testing each primitive type 44. **How does Backface Culling work?** - faces where the face normal points away from the camera are not rendered --- ## 03 - Application Areas and Application Development 45. **Name industries where AR can become a use case or change the daily work** - Medicine - Military - Product Presentation - Manufacturing and Maintenance - Logistics - Manuals - Automotive - Annotatins - Architecture - Art and Entertainment 45. **Name examples for medicine use cases** - Vein Visualisation - Diagnosis and treatment - Virtual Liver Surgery Planning System - Training of staff 45. **Name examples for military use cases** - simulations - training - Helmet mounted sights - X-Ray vision of buildings 46. **Name examples for product presentation use cases** - lego - cars - future products - magazines - games - catalogues 47. **Name examples for manufacturing and maintenance use cases** - error reduction - time reduction - remote support - manuals - assembly step tutorials - training 48. **Name examples for logistics use cases** - delivery - communication - picking - quality assurance 49. **Name examples for automotive use cases** - HUDs - motorcycles 50. **Name examples for annotation use cases** - Info of real life objects - museums - geotagging 51. **Name examples for architecture use cases** - collaborative design platform - collaborative design platform protocol - virtual reality - augmented reality 52. **Name examples for art and entertainment** - virtual studios - special effects - facade AR - westworld experience - reality fighters - pokemon go ### Application Development 53. **Name the steps in the app dev process** - planning - development - testing - evaluation 54. **On what should you focus on especially from the user point of view?** - Navigation - Interaction - user feedback - beware of the target audience --- ## 04 - Tracking Foundations and Optical Tracking 55. **What is tracking?** - dynamic determination of spatial properties at runtime - measure position and orientation of a physical object in real life - for consistency between real world and virtual world 56. **What is tracking in AR used for?** - display of augmentations - interaction 57. **What can be tracked in AR applications?** - objects - features - camera position 58. **What can be tracked in VR applications?** - head - hand and fingers - eyes - torso 59. **What must an AR app address/fulfill?** - tracking - calibration - registration ### Characteristics of Tracking Technology 60. **Name the characteristics of tracking technology** - Physical phenomena - measure light, sound, gravity, ... - Measurement principle - signal strength, direction, time of flight - Measured geometric property - distances, angles - Sensor arrangement - sensor fusion (multisensors), sensor sync - Signal resources - Passive: magnetic field, light - active: acoustic, optic, radio wave - Measurement coordinates - global vs. local, absolute vs. relative - Degrees of freedom - DOF; 3 translations (XYZ), 3 rotations (head, pitch, roll) - workspace coverage - working volume of tracking system 1m^2 to global eg. GPS - spatial scan - outside-In (Sensors outside the tracked device) vs Inside-Out (Tracked device contains sensors) - static accuracy - position and orientation tracking of a fixed object - dynamic accuracy - measurement of a moving object - tracking latencies - length of tracking, fixed and variable latencies, prediction 61. **How can accuracy be increased?** - Kalman filter - Combination of prediction and correction 61. **What is the Kalman Filter?** - Sensor fusion / learning algorithm - Tries to get the exact state from a set of faulty set of measurements (often noise) - Finds the most optimum averaging factor for each consequent state 62. **Difference between fixed and variable latencies?** - fixed: sampling sensors, pose estimation - variable: network communication, buffering, synchronization ### Registration 63. **What are the common AR registration issues?** - static errors (viewpoint and environment remain still) - Causes are Optical distortion, errors in tracking system and incorrect viewing parameters - dynamic errors (viewpoint and/or environment move) - Causes are Delays and lags - jitter (even with markers, objects could move w/o movement of them happening) 63. **Solution for jitter?** - If 3rd measurement is in range of 1st measurement, interpolation of 2nd could be performed by using 1st and 3rd measurement as base ### Tracking Systems 64. **Name different types of tracking systems** - GPS - Inertial tracking - Inertial sensing - Inertialsensing (accelerometer) - accelerometer and gyro - optical tracking - infrared light - beam scanning - videometric - inside-out - outside-in - pattern recognition - optical tracking w matrix code - spatial scan - SLAM (Simultaneous Localisation and Mapping) 65. **Explain Inertial tracking** - 2 components (accelerometer, gyro) - advantages: no stationary tracking, large indoor areas - disadvantages: positioning errors, orientation problems 66. **Explain Inertial sensing** - Information about relative transformation of a target - conserves axis of rotation - mechanical gyro 67. **Explain Inertial sensing - accelerometer** - measuring linear acceleration of an object - measures force on a mass (example: Spring with mass) 68. **Explain Accelerometer and gyro in combination** - acceleration and orientation measurement - measurements always relative to last position - Beware of accumulative error 69. **Explain GPS** - Global Positioning System - signals from three to four satellites - triangulation of three spheres 70. **Explain Optical tracking** - different types of markers (active, passive (reflect light)) - image processing is used - landmark recognition - inside out vs outside in - Infrared light possible as well 71. **Explain Infrared use of tracking** - IR camera to track markers - errors can occur like light source collision or bright surroundings 72. **Explain Inside Out tracking** - The tracked device itself contains the cameras or the sensory equipment required for tracking ### Spatial Scan 73. **Explain Videometric** - tracking of a users helmet - with mounted cameras at the ceiling 74. **Explain Beam Scanning** - Scanning optical beams on a reference - sensors on target detect time of sweep of the beams 75. **Explain Outside-In** - Cameras or sensors are placed outside the tracked device - High speed cameras 76. **Explain multiscopy** - two or more cameras - triangulation to get spatial position - multi-object measurement = orientation detection 77. **Explain Pattern Recognition** - one camera - shape and size of object detected (known from the system) 78. **Explain Optical Tracking - Matrix Code** - barcode - matrix markers attached on RL object - = object and pattern recognition - measurement and position calculation via the markers and a camera - Steps: 1. Binarisation 2. Connected component analysis 3. Coordinate System estimator 79. **What is SLAM?** - Simultaneous Localisation and Mapping - Idea: place a robot in a room and let the robot build a map of the room - using map to calculate the robot's location in the room 79. **What is odometry?** - The use of data from the movement of actuators to estimate change in position over time 80. **What parts are included in SLAM?** - Landmark extraction (object recognition) - Data association (matching observed landmarks from different scans) - State estimation (stochastic map building (Kalman Filter used)) - State update (position estimation + displacement calculation) - Landmark update - Application areas (AR tracking, robots [iRobot]) --- ## 05 - Mobile Augmented Reality and Head-Mounted Displays 81. **What are the components of a mobile AR system?** - HW computational platform - Display - Tracking - WIFI - wearable input and interaction - software 82. **Name devices used with AR** - backpack setups - smartphones - tablets - consoles 83. **What is used from a mobile device to track the device's location?** - Communication technology - WPAN (Bluetooth) - WLAN - WWAN (GSM and UMTS, LTE) 84. **Name the main challenges of mobile AR** - limited computational resources - size, weight - battery size - tracking and registration - 3d graphics - real time performance - networking 85. **Name software for mobile AR** - OpenGL - ARKit - ARCore - mixare - AndAR 86. **Name products of Head Mounted displays commonly known** - HoloLens - Google Glass - Google CardBoard - Samsung GearVR - HTC Vive - PlayStation VR - Oculus Rift 86. **Name binocular cues** - Accommodation, change of lens to bring object at certain depth in focus - Vergence, movement of eyes to map an object on the corresponding side of the retina - Binocular disparity, the difference in image location of an object seen by the left and right eye 86. **What is the Screen Door effect?** - It looks like you’re viewing the world through a mesh screen, and is a result of the black, empty spaces between pixels when seen up close. 87. **How do optical see-through HMDs work?** - mostly used in AR - similar approach as Head up displays in cars - real world light reduction - blending virtual objects in - optical blending requires no additional computing power 88. **How do video see-through HMDs work?** - recording real world - combining real with virtual world - presenting on a "monitor" - no depth info - separate video streams for RL and virtual world - resolution is limited 89. **Compare video and optical see-through hdms** - Advantages of optical blending - simple - computational cheaper than video blending - Field of view not a major issue - distortion needs to be compensated - display of real world is not degraded - direct view of real world still visible in case of power outage - no eye offset like on video blending - Advantages of video blending - fully obscured real world - composition on pixel-by-pixel basis - no semi-transparent systems - easier to match brightness of RL and virtual objects - RL and virtual view delays can be matched - problem with brightness (washed out real world or virtual world) 90. **Compare smart devices, smart glasses and hdms** - smart devices - intuitive interaction - have to be held - mainly monoscopic - smart glasses - low weight - reduced interaction possibilities - very low FoV - HDMs - hands-free interaction - encumbrance of the user - low FoV 91. **What is a Field of View?** - Field of View is the area the user can see through a HDM, AR Smartphone/Tablet or smart glass 91. **What different form of display technology exist?** - Prism lenses - WaveGuide Arrays - Half silvered mirrors --- ## 06 - Spatial Augmented Reality and Tangible User Interfaces 92. **What is Spatial Augmented Reality?** - Actual objects or real world surfaces are used as display area 92. **What is part of spatial AR?** - Augmented Surfaces - projection mapping - industry applications - virtual sandbox 93. **What do tangible user interfaces allow?** - more users - co-located interaction - encumbering one user 94. **What can be used as user interface?** - actual object surface - RL world surface 95. **What are the advantages of tangible user interfaces and the real world displays?** - display surfaces may not be just walls - flexible projection config - camera for calibration and tracking - rendering using texture mapping hardware - projector = fixed location 96. **Main problems with projector based tangible user interfaces** - gemoetric correction - color correction - intensity correction - display on arbitrary surfaces ### Augmented Surfaces 97. **What are Augmented surfaces?** - using real world objects as extension for displaying information - like table or wall - interaction must be possible 98. **What does the term hyperdragging mean?** - spatial dragging of virtual objects - like drag'n' drop - dragging from one display to another (other display is RL object) - establishing a link between RL object and virtual object 99. **What components do we need for augmented surfaces?** - Anchor cursor (link between RL object and virtual object) - Object aura (area around physical object) 10. **What do we need and need to do to add virtual objects to a physical object?** - A link between physical and virtual object is needed - area around physical object is needed - data attaching to physical object - binding virtual object to aura of physical object 10. **What do we need for Augmented surface setups?** - communication via WIFI - projection on info table or wall - identification of physical objects via markers and cam - additional camera (tracking) 1. **What is projection mapping?** - projection on surface or real life object like a house - calibration is needed - Complex calibrations corrections with help of projector camera (Procam) - for advertisement or art projects 1. **Explain the term AR sandbox** - kinect - projector - sandbox - Depth images are captured - heat map generation 1. **Name other augmented surface projects** - Augmented climbing wall - pico projectors - mono-on-mono - SixthSense and Wear Ur World ### Tangible User Interfaces 1. **What is a tangible user interface?** - physical object representing an object in the virtual world - may have physical controllers attached - TUI provides access to virtual information - how: intuitive physical manipulation - try to integrate UI in surrounding world 1. **On what objects can a TUI be attached to?** - solid - liquid - gases 1. **Name projects with TUI** - TUISTER - Toolstone - marker based input - iOrb - magnifing lens - Keyers - ChairIO - ShoeSoleSense 1. **How to build a TUI?** - Potentiometer - Arduino board --- ## 07 - Interaction 1. **What is interaction in general?** - influence of two entities on each other - interaction with environment occurs if user input is responded by the system - action --> reaction 1. **What two types of interaction exist?** - Abstract interaction (e.g. gesture using data gloves) - natural interaction (direct connection; virtual object pick up) 1. **What other ways of interaction exist?** - egocentric interaction - exocentric interaction 1. **What are the differences between AR and traditional desktops?** - no mouse or keyboard - 2D input not ideal for 3D object manipulation - multi-touch - speech/gesture recognition 1. **What is marker based interaction?** - use markers as pointing devices - e.g. state change mechanisms - visual representation by icons - buttons/sliders - Problems: - The action which can be performed is not always obvious to the user - Too many markers can be confusing 1. **How can marker based interaction be used?** - shake - rotation - hiding/showing 1. **What can be visualized with marker based interaction?** - Menus - Arrows - Symbols - Texts - Icons 1. **Tell the difference between manipulative and semaphoric gestures** - manipulative: couple target with gesture - semaphoric: symbolic/representation of static or dynamic gestures ### Motion Capturing 1. **What different types of motion capturing technologies exist?** - markerless (Kinect, leap motion) - simple markers (commercial systems) - bodysuit (VR domain) 1. **What is Motion Capturing used for?** - used for skeletal animation of virtual characters 1. **Explain hand tracking** - Use of LEDs, reflectors or color markers (FTIP) - data gloves - collision detected by distance between finger position and virtual object - multi-finger-detection: grabbing usable - problem: detection of multiple hands 1. **Tell me something about gesture tracking** - good for SAR and HMDs - HoloLens, Meta, Leap Motion - no defined starting point (problematic) - Can lead to fatigue (weight of hand etc.) 1. **Explain head tracking** - detection of nodding and head shaking - knowledge about head position and orientation 1. **Explain body tracking** - used for recognizing bowing down, sitting - e.g. Kinect ### Gaze based interaction 1. **What is gaze based interaction?** - common in VR - Eye tracking - gaze direction can be calculated by head transformation and eye tracking 1. **What is used if eye tracking not available?** - approximation of gaze vector ### Speech based interaction 1. **Explain Speech based interaction** - Good because hands-free - complex sentences distract user from detailed scene - avoid speech based interaction on heavy visual load 1. **Name the different types of speech based interaction** - Commandos - numerical input - free recognition 1. **What approaches exist?** - Commands + numerical input - interpreting free recognition - e.g. sphinx, julius ### Interaction Techniques 1. **Explain the Canonical 3D manipulation tasks** - Selection –> the task of manually identifying an object - Position –> the task of positioning an object from an initial to a final, terminal, position - Rotation –> the task of rotating an object from an initial to a final orientation 1. **Explain interaction processes according to bowman** - split up into three main tasks, Selection, Manipulation and Release - main tasks consist of sub-tasks - manipulation requires selection - selection does not always lead to manipulation - combine above mentioned techniques to create new ones 1. **Name the manipulation methods** - Direct user control (interface gestures mimicking real world interaction - physical user control (objects user can physically touch) - virtual control (devices a user can virtually control) 1. **Which variables affect the manipulation?** - distance to the object - object size - translation distance - amount of rotation - object density 1. **On what do manipulation techniques depend?** - interaction tasks - variables 1. **Explain the interaction technique Ray-casting** - = selection technique - ray cast from observer position in scene - collision between ray and interaction object is calculated - collision detection only in specific areas - after selection give feedback to user 1. **Explain the interaction technique Aperture** - = selection technique using hed and hand sensor - selection cone is cast into scene - depends on distance between hand and head size 1. **Explain the image plane techniques (an interaction technique)** - use of hand, finger, head tracking - Headcrusher (ray between head and center of an area) - sticky finger (ray between head sensor and finger sensor) - lifting palm (open palm below object) - framing hands (palm and thumbs in corner of frame) 1. **Explain the hand tracking technique jog dial** - fingertip recognition - fingertip orientation tracking 1. **Explain the hand tracking technique jog dial + slider** - fingertip recognition - rotation - slider recognition (sliding the fingertips) 1. **Explain the interaction technique "Studierstubes PIP (Personal Interaction Panel)"** - interaction via tablet and pen tracked - tablet/pen = tangible interface - asymmetric interaction (dominant hand) 1. **Explain the interaction technique Pick-and-Drop** - natural extension of drag and drop - stylus used - network communication - switch objects from one device to another ### Visual Hints 1. **What components are used for visual hints?** - tangible tooltips (on object recognition tooltip appears) - tangible bubble (like tooltip but a speech bubble) - ghosted hints (what action you can do with the object) 1. **What types of hints exist?** - textual hints - diagrammatic hints - ghosted hints - animated hints - placing annotations in the environment - Hedgehog labelling ### Menus 1. **How can a menu be represented?** - textual - graphically - 3D Objects - WIMP interfaces 1. **What are menus used for?** - changing states - adjusting scalar values - choosing from objects/options 1. **How can a menu be placed?** - world referenced - object referenced - head referenced - body referenced - device referenced --- ## 08 - Advanced Rendering 1. **What does advanced rendering do?** - issue to map real world data in virtual world data together - ideally in real time - mapping virtual content correctly 1. **What needs to be taken into account for real world to virtual world mapping?** - consistent lighting - use of shadows - realness factor of virtual experiences - virtual people 1. **What two aspects of the above question can be taken care of in AR?** - consistent lighting - use of shadows 1. **Why can they be an issue?** - just partial known environment - occlusion between real and virtual environment - lighting conditions need to match ### Stereoscopy 1. **What is stereoscopy vision?** - 120° FoV - binocular cues 1. **Name the binocular cues** - Accommodation (change of lens to bring object in focus) - vergence (movement of eyes to map object on the correspinding side of the retina) - binocular disparity (objects with different distance from eye focus) 1. **Name the issues for binocular cues** - accomodation-vergence conflict - Focal plane is always on the display surface - Accommodation is always set to get the focal plane sharp - Vergence has to change due to disparity - In real life accommodation and vergence always match - When using a display they can’t match, unless the display provides different depth layers - binocular rivalry ( perception alternates between different images presented to each eye) - diplopia - suppression 1. **Name the types of perspective projection** - parallel (two cameras with off-set) - toed-in (for each eye a camera, pointed towards single focal point) - off-axis 1. **Name the techniques of rendering stereoscopic images** - Temporal (Shutter glasses) - spatial (HMD) - color (Anaglyph) - autostereoscopic displays (barrier / lenticular displays) 1. **Name the relevant topics for stereoscopy in AR** - HMDs - depth perception - autosterescopic tablets ### Depth Perception 1. **What depth perception can occur?** - oculomotor cues (binocular display) - monocular (induced by motion) 1. **Name additional depth cues** - Depth of field - Motion parallax - Occlusion - Shadows - Shading 1. **Name examples for monocular cues** - relative size - relative height - familiar size - texture gradient 1. **Name examples of motion induced monocular cues** - deletion and accretion - motion parallax ### Distance Perception 1. **Why are distances underestimated by users in a virtual environment?** - measurement methods - technical factors - human factors - environmental variables - personal variables ### Visual Coherence 1. **What is a visual coherence?** - consistent environment between virtual and real content - seamless blending of virtual objects in real environment - easier to handle with video based approach 2. **What are problems of Visual Coherence?** - Occlusion - Lighting and Shading - Shadows ### Occlusion 1. **When does occlusion happen?** - if virtual objects are in front of real objects - if virtual objects are behind real objects - one of the strongest depth cues and should be resolved 1. **What is one of the main challenge of AR?** - correct occlusion between real and virtual objects - because no info about real world coordinates available from the beginning 1. **Name approaches which solve this issue** - modeling real world beforehand - generating depth maps - use of optical flow - real world object detection and extraction 1. **What is phantom rendering?** - phantom = virtual object represented as real object - phantom rendered invisibly only the Z-Buffer is modified 1. **How is phantom rendering done?** - draw image on color buffer - disable writing to color buffer - render phantoms on real scene in z-buffer - enable writing to the color buffer - draw virtual objects 1. **Explain the model free occlusion** - use of depth sensors to generate depth map - virtual content and real content merged in z-buffer - real environment is not modeled beforehand - can be freely modeled based on sensor input - Approaches: special HW or assumptions 1. **What are the sources of inaccuracy?** - phantom models not 100% accurate - errors in static registration - errors in dynamic registration - correction can be done in image space (post processing) 1. **How can occlusion refinement be done?** - edge detection - Camera has edge detection - AR application searches for edges - Edge smoothing - Polygon renderer displays rendered results - Others compute optical flow between camera image and model 1. **What is probabilistic occlusion?** - assumptions about the real world model - tracking required for rough representation - transparency increases from inner to outer models ### Lighting 1. **What is photometric registration?** - consistent illumination between real and virtual objects - knowledge of illumination of real environment is required - remote light sources and local illumination are easier to compute - consider reflection, refraction and shadows for rendering 1. **Explain image based lighting** - HDR maps used - representing illumination in physical units - radiance & irradiance map 1. **What is a radiance map?** - environment map representing light from the point of the observer 1. **What is an irradiance map?** - environment map representing outgoing light after reflection 1. **Explain light probes** - way to physically record radiance maps 1. **Explain the differences between active and passive light probes** - Passive: - reflective sphere placed to capture real world (300° FoV) - determination of light sources - Active: - placing a camera in the scene - record light conditions - fisheye, 360° camera 1. **Explain the term offline light capturing** - pre-processing light sources - multiple images can be used to generate a radiance map 1. **When do we need photometric registration?** - light probes too complicated - offline capturing too complex - instead using regular camera frames 1. **From where can we use data for photometric registration?** - static images - Help: user input - specular reflections - Reflection in human eye - diffuse reflections - Spherical harmonics - Face detection with machine learning - shadows - Knowledge of casting geometry has to be known - Has to be classified correctly - outdoor - Position and time for sun position ### Shadows 1. **What do shadows add to the scene?** - level of realism to rendered image - visual clues to determine spatial relationships 1. **Name the requirements for shadows** - real objects have corresponding 3D geometry (phantom model) - known light source - 4 different configurations available 1. **How are shadows displayed?** - shadow volumes (project a ray from the light source through each vertex of a shadow casting object into infinity) - shadow maps (Scene is rendered from the view of the light source) ### Diminished Reality 1. **What does diminished reality mean?** - taking objects out of the real world - by superimposing graphics 1. **Which three problems address diminished reality?** - determination of region of interest (dynamic set of pixels) - observation of hidden areas (view of background behind the ROI has to be integrated) - new content for removed content 1. **Tell me something about the region of intereset** - Region of interest is a dynamic set pixels on the screen to be replaced - It can be freely visible or partially or fully occluded - User may outline ROI, use strokes, selection rectangle - Object behind ROI can be used to define ROI - ROI could also be an articulated object which has to be tracked to be removed ### Stylized Rendering 1. **What is stylized rendering?** - emphasizing regions / aspects of the scene - - can be used for visualisation of art projects 1. **Which types of emphasizing exist?** - cartoon shading - pencil sketch shading - painterly rendering --- ## 09 - Collaborative Augmented Reality 1. **What is CSCW?** - Computer supported cooperative work 1. **Name problems of CSCW** - Disparity between people who work and the ones who get the benefit - Breakdown of intuitive decision making - seams (spatial, temporal functional constraints) - major categories of seams: - functional - Discontinuities between different functional workspaces - Seams between shared and personal space - Force the user to change modes of operation - cognitive - disruption between existing and new work practices - Abandoning acquired skills and learning new skills - Forcing the user to learn new ways of working - Often rejected if users have to change the way they work 1. **To minimize seams CSCW should support what?** - support of existing tools - support of existing working techniques - visual/audio communication between participants - Collaborators must be able to maintain eye contact and gaze awareness - Users must be able to bring real world objects into the interface 1. **Which types of collaboration spaces exist?** - interpersonal space (video conferences) - shared workspace 1. **For collaborative AR - which setup possibilities do you know?** - co-located (same physical space) - geographically dislocated 1. **What are the issues with VR HW concerning CSCW?** - images rendered for one viewpoint - human has to be rendered as avatar - low portability 1. **Advantages of AR in CSCW?** - users can see each other's facial expression - reference of real objects - no full rendering of environment needed - real users and virtual world on one display 1. **Name the five key advantages of collaborative AR environments compared to traditional VR setups** - Virtuality - augmentation - cooperation - independence - individuality 1. **Name aspects of collaborative AR users?** - typically co-located - no issues with viewpoint control - private and public space sharing has to be considered 1. **Name aspects of collaborative AV and VR users?** - typically dis-located - users immersed in VE - viewpoint problem - VE should replace real world 1. **Advantages of AR collaboration?** - shared virtual content - perception and sharing of stereoscopic 3D data - seperation between public and private space - inclusion of tangible, real-world objects 1. **Issues of AR collaboration?** - small FOV 1. **Name different coll. AR Applications** - TeamWorkStation - Translucent overlay of live video images of computer screens, two cameras for faces, private + public space - ClearBoard - Projection of video stream of remote user directly on the display - AR^2Hockey - Mixed Reality Stage - Projection of virtual characters and models in a real down-scaled environment - Transvision - Visualisation of CAD data in a shared environment - cAR/PE! - WearCom - Collaborative Web Space - 3D web browser, commands via voice - blue-c - Empathy Glasses - VITA - offsite visualisation of an archaeological dig, combination of various displays - Display in WIM and life size - AR Tennis - Co-located communication via Bluetooth, multi modal - CheckMate - HoloPortation 1. **What are awareness cues of AR Collaboration?** - users have to perceive remote users action - view frustum visualisation - gaze vector visualisation - potential issue: visual clutter 1. **Name awareness cues of AR Collaboration?** - Empathy Glass 1. **What are hybrid solutions?** - simular issues to Networked Virtual Environments - real objects hard to transfer into virtual space - AR users represented in VR/ VR awareness parameter visible in AR - General problem of awareness parameters is visual clutter - prototypes make us of hand presentation and eye tracking 1. **Networked Virtual Environments?** - Researches in: - Scalability - Responsiveness - Collaboration - Architectures - Communication protocols - Human factors - Provide useful and interesting asset to coll. AR 1. **Design Guidelines for Collaboration in MR?** - as much information of the remote environment - 3D Mesh of environment - updated 3D Mesh in realtime - provide independent PoV - as much Awareness cues as possible - transmitting speech - information of posture - poiting of hand should be modelled - cues for events outside due FOV - possibility to turn awareness cues on and off - usability and comfort - comfortable interface --- ## 10 - Evaluation Basics 1. **Evaluation in general** - important for app usability - performed by experts - three groups preferred (HCI experts, developers, users) 1. **What are typical problems for evaluation?** - Evaluation and user studies are one of the main topics in the humanities community - developers typically don't come from humanities community - demand for guidelines is high 1. **What are typical approaches of evaluation?** - expert studies - empirical studies - questionnaires during and after use - observation of the test subjects during interaction 1. **When and what data should be collected for evaluation?** - beforehand: - goal of a study - setup of tests - number and nature of tasks to complete - during evaluation - completion time of tasks - afterwards: - statistical evaluation - graphical evaluation 1. **What possible uses of psycho-physiological and neurophysiolocig measurements?** - estimation of mental workload - stress - strain - level of cognitive performance - alertness - arousal 1. **How can these measurements be performed?** - Electroencephalography (EEG) - Event-related potentials (ERP) - Electrocardiography (ECG) - Skin conductance response (SCR) 1. **Which statistical analysis methods/tools are used?** - Arithmetic mean - Median - Frequency distribution - T-test - Bonferroni - ANalysis Of VAriance - Variance - Standard Deviation - Quantiles, quartiles and percentiles - Displaying data: charts 1. **What data should be gathered and prepared for evaluation?** - amount of participants - gender - age - profession - degree of experience - forms - schedule - test plan 2. **How to evaluate MR apps?** - investigate human activity - participant observation - in situ interviews - user recordings (video) - collaborative viewing of recordings of a group - transcription - use of the three design loops: - Loop1: - User requirements - Technology opportunities - Selection and evaluation of solutions - Loop2: - usability studies (not always needed) - proof of concept - use of virtual prototyping - Loop3: - ergonomic research - general evaluation 1. **Examples of MR App evaluations?** - Mixed Reality stage - Lighthouse Trial 1. **What are the development domains of MR App evaluations?** - Behavioural domain - user centered approach - Interaction with the application - Usability of the user interface - Application content - constructional domain - Api development - every domain has experts 1. **Whats the 4 step sequential analysis?** - User task analysis - Expert guidelines-based evaluation - Formative user-centred evaluation - Summative comparative evaluations 1. **Standards for Questionnaires?** - NAS Task Load Index - Subjective assessment tool to rate the perceived workload of the user - System Usability Scale 1. **Aspects of evaluating with wearable computing applications using a HMD?** - for outdoor applications - main issue: supervisor is not able to see what the subject sees - interconnection of HMD via WIFI to supervisor's laptop 1. **Problems with AR applications?** - Latency - Depth perception - Adaptation - Fatigue and eye strain