Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

13
Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar

Transcript of Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Page 1: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Virtual Image PeepholeBy

Kyle PatienceSupervisor: Reg Dodds

Co Supervisor: Mehrdad Ghaziasgar

Page 2: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Quick Recap

• Development of an interface that does the following: • A very large virtual screen is

imagined to exist, • The mobile phone screen is used

as a "peephole" into that interface.

Page 3: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Overview

•Design Decisions and System Changes• Implementation •Finding and Tracking Features•User Interface •Tools

Page 4: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Design Decisions and System Changes

PREVIOSLY CURRENTLY

• No NDK • Uses NDK

• Use Android camera API • Uses device’s camera natively

• Didn’t use sensor readings • Uses both accelerometer and gyroscope

• Display was modifies camera frames • 3D interface

Page 5: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

NDK

Implementation

Get frames Process frames

Get sensor readings

Output

Page 6: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Finding and Tracking Features• The Shi-Tomasi Corner Detector

• Based on the Harris Corner Detector. • Tries to find little patches of image that generate a large variation when moved around.• Ultimately finds small corners in a frame.

• Lucas & Kanade Method• Technique which can provide an estimate of the movement of certain features in successive images of a

scene.• Algorithm makes a "best guess" of the displacement of a neighbourhood by looking at changes in pixel

intensity.• These intensities are known through intensity gradients of the image in that neighbourhood.

Page 7: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

User Interface using camera framesMoving the device without tilt will move the interface

Page 8: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

User Interface using gyroscope

Tilting phone to the left Tilting phone to the right

Page 9: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

User Interface using accelerometer

Tilting phone upwards

Tilting phone downwards

Page 10: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

ToolsPlatform Windows 8.1 x64

Applications Android Studio

SDK Android SDKOpenCV Android SDKAndroid NDK

Libraries OpenCVjPCT

Languages JavaC++

Page 11: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Project Plan

Term 1 Learn OpenCV Learn Android

Term 2Learn NDKBuild Android PrototypeCapture Footage

Term 3Process FramesProcess Accelerometer and Gyroscope ReadingsCreate 3D InterfaceIntergrade all components

Term 4ImplementingTuningTesting

Page 12: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

References

• Ali, S. I., Jain, S., Lal, B., & Sharma, N. (2012). A framework for modelling and designing of intelligent and adaptive interfaces for human computer interaction. International Journal of Applied Information Systems (IJAIS) Volume.

• Anuar, A., Saipullah, K. M., Ismail, N. A., & Soo, Y. (2011). OpenCV based real-time video processing using android smartphone. International Journal of Computer Technology and Electronics Engineering (IJCTEE), 1(3).

• Shi, J., & Tomasi, C. (1994, June). Good features to track. In Computer Vision and Pattern Recognition, 1994. Proceedings CVPR'94., 1994 IEEE Computer Society Conference on (pp. 593-600). IEEE.

• Tomasi, C., & Kanade, T. (1991). Detection and tracking of point features. Pittsburgh: School of Computer Science, Carnegie Mellon Univ..

Page 13: Virtual Image Peephole By Kyle Patience Supervisor: Reg Dodds Co Supervisor: Mehrdad Ghaziasgar.

Questions?