Problem Statement: In this assignment you have to implement Kalman tracking for multiple points, simultaneously, through a video. The code will take a video (either as one video file or as individual frames) and output the tracks of the points through the video.
Use Harris points in the first frame as the set of points to be tracked. Limit the number of points to be the 50 strongest point responses. Many of these points might not fall on moving objects, which is fine. The tracks generated for stationary points should be of zero length. Your tracking should be agnostic to which features are from the moving objects and which are not. It should be able to track objects that stop and restart.
Study the effect on the quality of the tracks with (a) state representation, i.e., location only,
location + velocity, and location + velocity + acceleration, (b) change in state model covariance, and (c) change in measurement covariance. Investigate at least 5 different variations for (b) and (c)
As output, overlay the entire tracks on the last frame of the video and show them as one image per video in your report. Show results on at least 5 videos of at least 200 frames, each. You can link to any video visualizations that you might have as well. Along with the track include visualization of the uncertainties involved using error ellipses.
You may use any video data available at https://motchallenge.net/ for your assignment.
Submission Requirements:
Please upload a ZIP file containing the following files:
1. All your code files, including any helper files/dependencies.
2. A README file detailing how to run your code along with any compilation instructions.
3. A 2-Page technical report containing the following sections:
a. A short description of the algorithm
b. A description of any code/algorithms that were used/re-used by you for your
implementation.
c. A few examples of results from your implementations, comparison with the original implementation (if needed).
d. A general discussion of lessons learned based on your experiments with the algorithm. E.g. What did you struggle with, issues faced while implementing the code, scopes for and/or proposed improvements, etc.
Grading:
Each assignment will be graded out of 100:
Code (out of 30) quality of coding, readability, understandability (comments, variable names, etc.)
Report (out of 40), point 3a, 3b, 3c, and 3d above
Demo and Results (out of 30): Compilations, execution, correctness of results on test cases.
Solutions to your programming assignments have to be self-sufficient and not dependent on other computer vision codes, such as OpenCV or Matlab vision package. You may use packages for display graphics or mathematics packages, such as for linear algebra (NumPy, for example, Matlab (but not computer vision module)) or graphs or optimization. You can use any code distributed in the class.
Codersarts is a top rated website for students which is looking for online Programming Assignment Help, Homework help, Coursework Help in C,C++,Java,Python,Database,Data structure, Algorithms ,Final year project,Android,Web,C sharp, ASP NET to students at all levels whether it is school, college and university level Coursework Help or Real time project. Hire us and Get your projects done by computer science expert.
Comments