In the last meeting I firstly showed the video of tracking algorithm using FERN descriptor as well as two most recent descriptor, namely ORB and AKAZE. Followed by that, Qiang found two footages (1), (2) during the meeting and he asked me to try out these video or something similar to them using our developed algorithm and applied one of these feature-based tracking as a proof-of-concept as well as for the purpose of experiment for future investigation. But during the meeting I was trying to explain the process of FERN but I started to get confused and thereby did not explain it properly...so I will also need to revisit and come up with a better and clear explanation.
- Revisit FERN descriptor and summarise the overall training process.
- Use the provided videos link as an input in order to examine the overall performance using our algorithm + current tracking technology.
- Revisit the current MEng project and start improving the fixing up the minor / major issues.
- DSP (See the sample)
-
This page is focused on the problem of detecting affine invariant features in arbitrary images and on the performance evaluation of region detectors/descriptors.
-
An Improved ORB Algorithm of Extracting and Matching Features
Research paper on improved ORB algorithm, 2012.
-
Pose Tracking from Natural Features on Mobile Phones
A research paper as suggested from its title.
-
General discussion fron Stackoverflow, on how BREIF works.
-
CV course from Princeton University
An archive for CV course from Princeton University.
-
Local Invariant Feature Detectors: A Survey
An overview of invariant interest point detectors, how they evolved over time, how they work, and what their respective strengths and weaknesses are.
-
A C++ library to provide SfM implementation, haven't personally try it yet but sounds interesting.
-
A set of useful computer vision functions by Peter Kovesi.
-
Lightweight library that provide machine learning and computer vision functions. I am currently using this as part of C++ implementation for master project.