COVID-19 AUTHORIZED ENTRY SYSTEM USING ADVANCED DETECTION TECHNIQUES:

OBJECTIVE:

• To reduce the spread of Covid -19 at crowded areas, thereby increasing safety at these public places.
• To provide a complete authorized entry system.
• To check for proper wearing of mask, temperature, and proper Spo2 level without any human intervention periodically.

INTRODUCTION:

• As normalcy is being slowly restored, during this covid-19 pandemic, it is crucial that the health and wellness of people are ensured.
• Instead of the conventional checking system which involves human intervention, facial recognition along with temperature and spo2 readings can be deployed to register the attendance and collect information about the people.

TRAINING MASK DETECTION MODEL:

DATASET USED : 

• Initially Custom dataset – 750+ Images with 3 Classes. (With Mask, Without Mask, Improper Mask)

TRIAL AND OUTCOME:

• Tensorflow Lite - Able to detect but was not accurate
• Yolo v3 - Computationally Heavy to run on an embedded system platform.
• Tiny Yolo - Able to detect but was not accurate
• Harr cascade (image processing) – Only able to detect face perfectly but not the mask.
• Tensorflow 2 - Computationally Heavy to run on an embedded system platform.
• Tensor RT - was able to detect with good accuracy and FPS.

Jetson Nano: Deep Learning Inference Benchmarks:

TRAINED MODEL PERFORMANCE:

SPO2 AND HEART RATE USING MAX30102:

• Prebuild Libraries for Max30102 was taken from GitHub and was programmed to get heart rate and spo2. [5]

Voice Recognition and Voice Command design:

• Speech recognition was carried out using the SpeechRecognition package in python.[13]

(was carried out in ubuntu on my laptop)

Using FireBase Database:

Hardware Setup:

Working video:

REFERENCES:-

1. https://developer.nvidia.com/embedded/jetson-nano-dl-inference-benchmarks
2. https://www.checkupnewsroom.com/7-types-of-thermometers-and-to-what-degree-you-should-trust-them/
3. https://www.kaggle.com/ashishjangra27/face-mask-12k-images-dataset
4. https://github.com/dusty-nv/jetson-inference/blob/master/docs/pytorch-collect-detection.md
5. https://github.com/vrano714/max30102-tutorial-raspberrypi
6. https://www.ti.com/lit/ds/symlink/lm35.pdf
7. https://www.microchip.com/en-us/product/MCP3008
8. https://scruss.com/blog/tag/lm35/
9. https://www.kaggle.com/andrewmvd/face-mask-detection
10. https://www.checkupnewsroom.com/7-types-of-thermometers-and-to-what-degree-you-should-trust-them/
11. https://developer.nvidia.com/embedded/jetson-nano-dl-inference-benchmarks
12. https://pythonbasics.org/python-play-sound/
13. https://www.thepythoncode.com/article/using-speech-recognition-to-convert-speech-to-text-python
14. https://www.c-sharpcorner.com/article/firebase-crud-operations-using-python/