CPSC 599 Biometric Technologies

Assignment 3 Requirements

Due: March 31st, Friday, 9:00 pm by e-mail to your TA

This is an individual assignment that intents to allow you to have a hands-on experience with various biometric recognition and synthesis software systems.

It has three parts: Part I, Part II and Part III. It is worth 20% of your course mark.

Part I Objectives GAIT

Part I of your assignment allows you to have a hands-on experience with gait recognition and 3D joint temporal data processing. This part of assignment should be performed in three steps:

Step 1.You will be provideda text file “traindata_labeled.txt” that contains temporal skeleton joint data from 5 different person. For each person, a total of 4 videos were recorded. The data was collected using a Kinect sensor and it contains 3D joint coordinates (x, y, and z values) of 20 different skeletal joints for each frame of the video sequence.

Each line of the text file represents an entire gait video sequence of a participant. Each line starts with a numeric value between 1 and 5, which represents the person (since there are 5 persons, we have used 1, 2, 3, 4, and 5 to represent them). The next integer F represents the number of frames in this current gait video sequence. Each of these frames contains the 3D coordinates of 20 skeletal joints. Hence the next 60 floating point values represent the x,y,z coordinate values of 20 skeletal joints in the first frame, the next 60 values are from the second frame, and so on.In other words, each line of the text file has the following format

person_label no_of_frames x(1,1) y(1,1) z(1,1) x(2,1) y(2,1) z(2,1) …. x(20,1) y(20,1) z(20,1) …. x(1,no_of_frames) y(1,no_of_frames) z(1,no_of_frames) …. x(20,no_of_frames) y(20,no_of_frames) z(20,no_of_frames)

*The values are separated by spaces.

Here, x(j,f) represents the x value of j-th joint in the f-th frame. Hence, each line sequentially represents the joint coordinate values of each frame.

The ordering of the joints is as follows:

1 HEAD

2 SHOULDER_CENTER

4 SHOULDER_LEFT

5 ELBOW_LEFT

6 WRIST_LEFT

7 HAND_LEFT

3 SHOULDER_RIGHT

8 ELBOW_RIGHT

9 WRIST_RIGHT

10 HAND_RIGHT

11 SPINE

12 HIP_CENTER

14 HIP_LEFT

15 KNEE_LEFT

16 ANKLE_LEFT

17 FOOT_LEFT

13 HIP_RIGHT

18 KNEE_RIGHT

19 ANKLE_RIGHT

20 FOOT_RIGHT

Step 2.Design a feature representation for each of the gait samples. You can try features like joint distances or angles formed by different joints, and look at how those values changes in each frame. Or you can look at their means and standard deviations. Propose a set of features that you will use to represent the gait sequences.

Step 3. You will be provided a text file “testdata_unlabeled” that will have some other gait recordings of the same 5 persons. In this case the data will be unlabeled, which means the person_label will not be provided. Your task would be to extract the same features from these samples and find a match in your traindata_labeled.txt file. You can use Euclidean distance to find out which sample in your labeled data has the closest distance with a test data.

The “testdata_unlabeled” contains 10 unlabeled recordings of the same persons (different recordings that what you will have in the traindata_labeled.txt file). Your task is to predict the labels for the test data.You should use MATLAB for this task, free student version of MATLAB is available at:

Submit a reportPART Ion your designed feature representation and your predicted class labels for the test data, not exceeding ONE PAGE in length. Also please submit your code asmt3part1_first_lastname.matfor this part of the assignment to your TA, separate from the report. Please contact your TA for any questions on this part of Asmt.

Part II Objectives SYNTHESIS

Part II of your assignment allows you to have a hands-on experience with fingerprint and face expression SYNTHESIS software:

  • FaceGen expression software
  • SFinGe fingerprint synthesis software

with links to free versions available at the course web site:

You must download and install both software systems and perform the following two experiments, and SUBMIT a report:

  1. Create TWO distinct synthetic fingerprints, one similar to an average fingerprint, one with highly unique features, and explain which settings in the software system you used to perform this task. Attach images of fingerprint to your report part II fingerprints. This portion of your report should not exceed one page.
  2. Create TWO synthetic faces, one similar to average (mean), one with highly unique features, and explain which settings in the software system you used to perform this task. Attach images of fingerprint to your report part II facial expressions. The portion of your report should not exceed one page.

Part III Objectives SOFTWARE USABILITY

Part III of your assignment allows you to compare features of some freely available software systems. You must download and install ONE new software system of your choice (you can find free download on the web site, or use site, for example).

SUBMIT a report:Please provide a one page TABLE which includes

Software name and link to download:

Description of purpose:

Key features:

Interface/Type of input:

Ease of use (0-5)

Sophistication (0-5)

Usefulness (0-5)

Portability (0-5)

Reliability (0-5)

All three parts must be submitted in a single word or pdf file, where code for part I also must be submitted separately. Please label code as Asmt3_FIRST_LASTNAME.

This assignment is to be done individually.