December 2011
1 post
FYP CA2
Introduction
Augmented Reality is defined as the modification of human visual perception of reality by imposing additional stimuli generated by the computer, such as 3D images. The stimuli can take many forms such as audio, visual markings, or even tactile feedback. As a comparison, virtual reality completely substitutes the real world with computer-generated reality.
There are many applications...
October 2011
2 posts
Week 8
After studying the EPnP source code and demo code last week, it is time to start hands-on coding. The aim is to integrate the EPnP code into the object-in-video that I made during the recess week, in order to obtain the estimated camera translation and rotation.
Calling the function compute_pose() triggers a break point when the program runs. This problem is rectified by keeping the number of...
Week 7
This week I have started on reading up the EPnP. According to the paper “Accurate Non-Iterative O(n) Solution to the PnP Problem”, the PnP problem is to determine the camera position given it’s intrinsic parameters and a set of n correspondence between 3D points in the real world and their 2D projections.
I am to study the EPnP source code and demo code. Basically, the code takes in Xw, Yw and...
September 2011
8 posts
WatchFeature-based detection using the OpenCV library. The window on the right displays the keypoints of the frames in grayscale. The left window shows the corresponding keypoints between the object and frame when detected.
Week 6
Started on video capturing. I had tried to run a code written on C++ for capturing the frames of a video, copy it and turn it into Mat (matrix class in OpenCV) format and display it on the monitor. However it could not be compiled. It took me 1.5 days to figure out that I failed to enter the cv namespace; I’ll never forget that again. Next step will be to combine the code I had learnt on week 5...
Recess week
Side note: on the 18th of September, there was a news report about an AR machine at Changi Airport Terminal 3 that looks similar to the “Magic Mirror” in the MXR lab. Its purpose is to allow customers to try on different clothes virtually on the screen, by dragging the virtual clothes onto the user.
This week, a big misunderstanding has been cleared. I had thought that the frames captured by the...
Week 5
This week is the start of hands-on coding. I have been given a sample code for detecting the most stable 2D points, i.e. the keypoints, of 2 images. One image is a picture of an object, and the other a picture of a scene with that same object in it. Bascially, the code compares the keypoints of both images in order to find the correspondence between the keypoints of the 2 images, and then map the...
Week 3
This week, I had borrowed some readings on AR to have a general idea on the types of AR, and the general idea of how it works. According to “Trends in Augmented Reality Tracking, Interaction and Display: A Review of Ten Years of ISMAR”, the core of AR can be broken down into 4 main research areas:
1. Tracking techniques
2. Interaction techniques
3. Calibration and...
Week 4
Due to my negligence and lack of time management, I was only able to start my FYP proper this week. I was briefed on the idea of this FYP about using AR for military training purposes, namely the CBRE.
I was tasked to read up on the Ferns algorithm. According to “Fast Keypoint Recognition usng Random Ferns”, it is an algorithm that finds the most stable points of an image through...
Timeline
Week Objective
3 Learn anout AR in general
4 Learn about Ferns algorithm for tracking
5 Understand code for tracking with Ferns algorithm
6 Combine video capture and tracking code
Recess week - Combine video capture and tracking...
FYP project aim
AR for military use. In AR, there are 2 areas of tracking and recognizing images: in 3D plane and in 2D plane. My area of research is in the 2D plane, specifically in 2D planar surface object detection.
