EyeAttend — Facial Recognition based Attendance System from scratch. — A complete approach. (Part-I)
— From Data Collection to Model Compilation to Backend Data Storage to Frontend Android App
The project would have been incomplete without the hardwork of my team members Abhijeet_Baruah and Aarushi Sood. We are thankful to our professor, Dr. Varun Gupta (HOD, Applied Sciences)for his esteemed guidance and motivation to keep us going and growing with this project.
Before we begin…
We would like to clarify that this won’t be one single story but a collection of stories as the project took a long time in development and with these stories, our main aim is to clarify each part of the development so that you are able to build similar, in fact more advanced version of this project for yourself.
Since the series feature is deprecated from Medium, the complete story will be divided into parts and the number will be mentioned in the title to keep a track of the post.
Motivation
The project is our final year’s mini project and we have chosen the domain to be Artificial Intelligence. Digging deep inside artificial intelligence lies the domain of Machine Learning and we have chosen our base to be Deep Learning (subset of ML).
We remember that during our first year of college, we felt a need of something inspiring, something that would amaze us and hence this thought became the motivation for our project. Following this motivation, we decided to give back what we thought at that time, almost after four years of B.E. in CSE via this project.
Our vision for this project is to introduce an intelligent system for motivating freshman year students in machine learning. The aim is to deploy this system in our college so that, once the freshman students observe its application in their own lives, this would definitely ignite the fire in some students to build a better version, better projects that can be used for socio-eco platforms.
The target is easy as every class has it’s own class representative and one of the main duty is taking the attendance of students. In this way we can have the direct interaction with the target audience i.e the fresh minds of college.
Takeaway from this series
- A scalable design with the focus on making the attendance system not only automatic but also intelligent.
- The architecture of the system mainly focuses on its usability and scalability which will easily help us to redesign it in accordance to the future feedbacks.
- With a modular approach, ML models can be plugg in/plug out.
- Use machine learning and vectorization for faster results.
- Android Application- “EyeAttend” for interacting with the system.
What will you learn to build
The video given below, will demonstrate the core working of the project. The foremost step is to select an image including all the people present in the room, sending it over to the server which will run the model on the image and return the list of presentees and absentees students. The project loads the attendance table for a particular teacher of a particular subject of a specific class in the database dynamically i.e creating tables, adding roll numbers, adding the date of attendance and also marks the attendance. The attendance can be downloaded for the whole class by the teacher and a student can download his own attendance for whole semester for a specific subject of a particular teacher, just in one click. The following demo shows the core working of the project.
Conclusion
It’s time to conclude the introductory story for this project. It will be our effort to provide the next part of the story once a week. We hope you will enjoy the above demo and information about our project and you are really excited to learn how to build the automated attendance system on your own. Let’s catch up in the next blog. Till then, have a great week! For getting updates about next part, follow us on medium, LinkedIn : Keshav Tangri, Abhijeet Baruah, & Aarushi Sood.
References
[1] Logo Design : https://www.designfreelogoonline.com/logoshop/free-logo-maker-cctv-camera-logo-design-copy/
