It gives me immense pleasure to announce the best project award for our online course “Computer Vision for Faces” (cv4faces) for this semester.
Why do we have a Final Project in our course?
We, humans, have an immense capacity to fool ourselves. We believe we can learn by reading, listening to lectures, or running someone else’s code. That’s all a good start, and I do not denigrate the importance of starting any which way. It is also ok to feel good in the beginning so we get a sense of progress.
But at some point, we have to commit to learning through action. As a kid, you probably learned to ride a bicycle with training wheels or with assistance from an adult. One fine day, the training wheels were taken off, and just like that, you went bicycling without fear.
Nah, let’s be honest. That never happened! Probably this is what happened. The training wheels came off and moments later you crashed and tasted the dust. You were possibly bruised and were made fun of by your friends.
If you know how to ride a bicycle today, you probably dusted yourself up and got back on your bicycle. You learned through pain and suffering. The learning you received through action permanently evolved your sense of balance, and soon you were able to ride without a thought.
The final project in our course is meant to symbolically take the training wheels off. We are there to help, but you are pretty much on your own. You decide the project topic, you formulate the solution, you try hard to make it work, you taste failure in the process. We help you if you get stuck, but we do not spoonfeed project ideas or solutions.
In the end if you have struggled hard with your final project, you will have a sense for how to solve real world problems, and you will undoubtedly sharpen your intuition.
Best Project Award
We launched cv4faces more than a year back and almost 850 students from more than 50 different countries have enrolled in the course so far. Most of the students who took the course did not have much exposure to Computer Vision, Machine Learning or AI. Yet, the students who were ready to put in the effort learned valuable lessons — some found new jobs and others found a new life.
For the best project award we considered all submissions between Oct 14, 2017 and Oct 14, 2018. Students were judged on Creativity, Quality, Difficulty and Effort.
$1000 in Prizes
We award a total of $1,000 in prizes. The first, second and third prizes are $500, $300, and $200 in cash respectively.
While looking into projects submitted in the course, we were amazed by the amount of diversity we had in our students in terms of their education, age, gender, and geography. We have students who are in high school and also who have completed high school 50 years back!
Let us now look at the winners!
Third Prize : FaceMask iOS App by Andy Liang
Andy is a C# developer and was interested in creating applications similar to SnapChat. Using the concepts covered in the course such as Facial Landmark detection, Delaunay triangulation, Warping and blending and some other external resources, he created an iOS app that takes the picture of a person and masks it onto the face of the person holding the mobile device, similar to snapchat. Given below are some sample results from the app.
As you can see the output is much less noisy than Snapchat and looks more realistic. We are very impressed with the results to say the least.
Second Prize : SpookFish – Face Recognition and Analysis App by Santhosh K.S
Santhosh is a Software Engineer. He took the course as he wanted to get into the CV-ML domain. He turned his learnings from the course, specifically the Deep Learning Face Recognition module, into a webapp. A video demo of the project is shown below.
What does it do?
It takes a youtube video and analyses the faces in each frame.
- For each frame, it detects all the faces and clusters them.
- You can enroll a person by giving it a name so that it can be recognized in subsequent frames or videos.
- You can also get the overall statistics of the video such as the number of faces in each frame, total number of people in the whole video, timestamps of appearance of a particular character etc.
Apart from this, more features can be easily added.
You can find the code for the project from the github repository.
We are really impressed by the amount of work put in for the project as it requires a lot of software engineering apart from the computer vision stuff.
First Prize : Coin Identification and Counting by Brian Tremaine
And the first prize goes to Brian Tremaine!
He has an engineering degree from Stanford University and is the owner of a circuit design and simulation consulting firm. He took the course to dive into this exciting field of Computer vision and AI. Combining several concepts taught in the course he created a solution to detect and count the coins spread over a flat surface.
The project requires a very good understanding of traditional image processing concepts such as blurring, edge detection, perspective correction, Hough transforms, etc. He also used camera calibration to improve accuracy. A sample output of the project is shown below.
Click here for the code.
While doing the research for this post, we found out that Brian is truly an inspiration for all of us. He is 69 and yet has the fire in the belly to not only learn new things but win competitions!
There’s a saying which aptly describes his spirit:
“It’s never too late to start”
If you liked this article, please subscribe to our newsletter. You will also receive a free Computer Vision Resource guide. In our newsletter, we share Computer Vision, Machine Learning and AI tutorials written in Python and C++ using OpenCV, Dlib, Keras, Tensorflow, CoreML, and Caffe.