Smart attendance system using image processing

The project aims to take attendance of students. Attendance is critical to a student's academic achievement in high school and college. Absence from class on a regular basis raises the chance of failing and early dropout. An attendance check is an effective method of increasing a student's atte

Project Title

Smart attendance system using image processing

Project Area of Specialization

Artificial Intelligence

Project Summary

The project aims to take attendance of students. Attendance is critical to a student's academic achievement in high school and college. Absence from class on a regular basis raises the chance of failing and early dropout. An attendance check is an effective method of increasing a student's attendance percentage in a class. The old procedures include calling names or signing documents, which are both inefficient and insecure.

Given that our object identification is based only on a person's face, it is preferable to store just the face portion of a person's picture, eliminating the rest of their body or surroundings. Thus, a face detection approach is necessary to automatically identify the region of the face in order to trim and store a picture consisting nearly entirely of the face for subsequent face identification.

Therefore, using the concept of neural networks and the "yolo" algorithm we have designed a smart attendance system.

Project Objectives

The objective of our project is the following:

• Count the number of students in a class or a close environment.
• Counting the number of people for surveillance on large scale.

Project Implementation Method

Custo Dataset was collected for training a model. After the collection of the dataset, we used roboflow to configure a bounding box across the area of interest, in our case "head". 

Roboflow distinguished the images given into training and testing images from our custom dataset.

After the images were collected we used google collab(an online GPU service) to train our model with the custom dataset.

After training our results are showcased on a laptop webcam.

Benefits of the Project

The project benefits in collecting student attendance eliminating the concept of proxy attendance. Moreover, on a broader scale, we can use the project for surveillance prospects.

Technical Details of Final Deliverable

Our project works on the Yolo v3 algorithm. We used Roboflow which separates the images from training and testing. Roboflow separated the images on basis of Yolo v3 and gave a file containing the images and coordinated the bounding box of each image. 

Once the file is received we uploaded the file on a google drive and mounted it on google collab(a free GPU service). Once the path is given we will call the file and the training starts. At first, we are given weight files of our custom dataset and then after that, we use them to train our model. 

Once the training is complete after several iterations our result will be displayed on an external webcam whose path we will give from google collab 

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Education

Other Industries

Others

Core Technology

Artificial Intelligence(AI)

Other Technologies

Others

Sustainable Development Goals

Quality Education, Reduced Inequality

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com