The hand traces database has 3 tracings for each person. You may either elect to work with each individual image, or with an average image of each person's hand. Java code for working with images is found in Lab 5.
Note. You may download the images of the hand traces database. by either clicking on each image and then saving it to your local machine, or by FTP'ing to dataprivacylab.org as you had done in Lab 5 and then downloading the images from the handtraces subdirectory found there.
Submit an abstract (up to 2 pages) describing your methods and statistical results. Send the abstract to paddataprivacylab.org by the due date for Project Assignment 1.
If you are taking this course for graduate credit, then complete at least 9 measurements for assignment 1.
Give a 5-minute presentation to the class on your experiment and findings for either this assignment (Assignment 1) or the next assignment (Assignment 2).
Note. You may complete assignment 1 and then later change your mind about which project you will in fact provide as your term project, provided your final decision occurs prior to the second project assignment and is approved by the instructor. See the course schedule.
Assignment 2.
For assignment 2, consider which set of measurements make scientific sense.
You may use those defined in Lab 2 or some other measurements of your own design.
Provide a confusion matrix based on your results from Lab 2 and Assignment 1
to determine why the measurements you have selected make sense.
The set must contain at least 9 measurements. We term this your
set of "sound measurements."
Run some experiments as needed to justify your set of measurements.
Submit an abstract (up to 2 pages) describing your choice of measurements and your algorithms for computing them. Send the abstract to paddataprivacylab.org by the due date for Project Assignment 2.
Final report. Modify your program from Assignment 1 to include your sound measurements. Given a trace of a hand, your program should compute at least 9 measurements that are scientifically informative and discrimminating.
Use your program on all the images in the hand traces database and determine which two images are closest for any two images. Report your findings. You may use any reasonable distance function.
If you are registered for graduate credit, then update your program so that given a trace of a hand and a set of known traces, your program will automatically determine which of the known hand images is closest using your sound measurements and a sound distance function. That is, you must justify why your distance function makes sense. Your program should return a rank ordering where the closest hand appears first, the second closest second, and so on.
Write a final report for the project and prepare and conduct an in-person poster presentation of your work.
Final report. Consider which set of measurements make scientific sense. You may use those defined in Lab 2 or some other measurements of your own design. Provide a confusion matrix based on your results from Lab 2 and Assignment 1 to determine why the measurements you have selected make sense. The set must contain at least 9 measurements. We term this your set of "sound measurements." Run some experiments as needed to justify your set of measurements.
Modify your program from Assignment 1 to report values for your sound measurements on the scanned images of hands.
If you are registered for graduate credit, then update your program so that given a scanned image of a hand and a set of known images, your program automatically computes the sound measurements and reports which other known hand is closest. You may use any resaonable distance function.
Write a final report for the project and prepare and conduct an in-person poster presentation of your work.