IoT Entry System Based on Deep Learning Recognition

2021 ELE Engineering Design Project (FM06)

---

Faculty Lab Coordinator

Farah Mohammadi

Topic Category

Consumer Products/Applications

Preamble

With the introduction of the 5G era, integrating deep learning models to IoT devices might bring about a new generation of apps capable of executing sophisticated sensing and recognition tasks, enabling a new world of interactions between humans and their physical environment. Convolutional Neural Networks (CNN) may be utilized effectively in mobile apps for this purpose if they can collect data (pictures and sounds) from a variety of IoT equipment, are energy-efficient, reliable, and can function with minimum data labeling. The goal of this project is to design and create an IoT image and speech recognition system based on Deep Learning (DL) as CNN to build facial images and voice recognition algorithms for mobile phone Apps.

Objective

Image and voice recognition are important in authentication procedures such as item identification, facial recognition, and speech detection. The objective of this work is to use an intelligent security system to control home automation system, so that the customers can used this application for multi tasks such as opening or closing doors using their voices and face image via the Phone App.

Partial Specifications

Iphone/Android APP Specification:
• Create a customized iPhone/Android App for smart identity recognition control system;
• Enable the App to communicate with multiple voices or facial images recognition units;
• This App have ability to register and initialize Multiple voice and faces image recognition units with password;
• The App have ability to receive a message from the recognition units when open/close door is happened;
• The App have ability for the history information per voice recognition unit on a daily, weekly, or monthly basis.
Smart identity recognition control system:
• The program should be able to provide connection between a user interface or DL platform with the Wi-Fi system;
• The Microcontroller software should receive control signal from DL platform or user interface and provide the required commands to perform automated household tasks such as opening or closing doors.

Partial Specifications:

-Architecting a CNN image and voice recognition model.
- Execution of DL classification on the proposed classifier architecture.
- Comparing the obtain results based on classification accuracy and select optimum CNN algorithms. The MATLAB or tensor flow platforms can be used to evaluate the proposed method.

Suggested Approach

1- Build a Deep Learning CNN platform as Graphical processing units-based tensor-flow, or MATLAB for training DL recognition systems.
2- Dataset Acquisition: Collect data sets for face images and voices for the learned classifier.
3- Use a Wi-Fi connection such as ESP8266 unit or similar devices, Arduino Uno microcontroller board, Raspberry Pi or any other similar microcontroller. Electromagnetic switch relay to control on doors opening.
4- Use appropriate open source for communication between a user interface as DL platform with Microcontroller and the cell Phone.
5- Use appropriate software and protocol to perform the WSN connection in real time via WSN to the router.

Group Responsibilities

1. Perform Literature review;
2. Collect and design the technical information to implement the system under consideration.
3. Implement and train DL techniques for image and voice recognition classification with the above objective.
4. Write a technical report and present the obtained results at the end of this project. Describing the main step of running CNN platform and showing the main simulations results.
5. Implement and verification of the entire system

Student A Responsibilities

• Build the CNN image and voice recognition models: Training the classifier for different DL models: as Single Conv2D Layer model, Two Conv2D Layer model or MobileNetV2 Model.
• Collecting training Data set (facial images and voices) and features extraction if possible.

Student B Responsibilities

• Design and implement Wi-Fi connections between microcontroller and user interface or DL platform.
• Design the required control circuit that provide the required commands to perform automated household tasks such as opening or closing doors.

Student C Responsibilities

• To design and perform the programing related with iPhone/Android App software

Student D Responsibilities

• To provide and design the total algorithms and flow chart and main working steps between the user interface or DL platform, microcontroller and cell Phone App.

Course Co-requisites

Digital Systems, Programming in C++, Microprocessors, python, MATLAB

To ALL EDP Students

Due to COVID-19 pandemic, in the event University is not open for in-class/in-lab activities during the Winter term, your EDP topic specifications, requirements, implementations, and assessment methods will be adjusted by your FLCs at their discretion.