Browser Based Application built during Microsoft Engage Mentorship Program 2022
Project Demo Link - https://drive.google.com/file/d/1Iz7IrT2musOWkGsx0pbTWj9Y7KIlKyME/view?usp=sharing
Table of Contents
To install and run the project on your local system, follow the given steps:
- Clone this repository
$ git clone https://github.com/riddhic15/safeHome.git
- Change directory to safeHome
$ cd safeHome
- Make sure you have python and pip are installed in your system. Do this with the following commands:
python --version
pip --version
If they are installed, their version will be displayed. To avoid errors in installing other libraries, upgrade your pip using the following command:
pip install --upgrade pip
Install all other dependencies that have been used in the project using pip:
pip install -r requirements.txt
- Run the app
flask run
NOTE: If you encounter an error saying module not found or an import error while importing app after running the 4th command, install the dependency that has lead to the error.
The prime objective of this project is to combat the constant fear of any mishappening or crime taking place at our homes in our absence. My application, safeHome, through its monitoring system singlehandedly guarantees 24x7 complete security of the home even when no one is present. So, no longer do we need get daunted to leave home even if we have to stay away for days or months.
The main goal of the project is to prevent crime and help people lead a safer and stressfree life. Face recognition allows the application to recognise and remember the faces of added members. Thus, it is able to detect unrecognised people by distinguishing them from the added people who may have broken into the home with fraudulent intentions. So, this recognition technology combined with emotion recognition result in success of this project.
Some of the features included in this app are:
Salient Features:
- Two modes of Operation : Home Mode and Away Mode
- Live Alert system through mail on detecting unrecognised faces
- Instantly add new faces
- Confidence Rating
Family members and friends can be added to the list of known people. After clicking add, about 100 pictures are captured which are then used to build dataset and train the model so as to recognise the people if they happen to be detected by the camera again. Names of people who have been recognised by the app, appear on the members page.
The confidence rating is a specialised feature developed to detect any unrecognised person whose intentions do not seem right on the basis of their facial expressions. This rating is calculated on the basis of whether the person is recognised or not and also on the basis of a person's facial expressions. So, any unrecognised person with fraudulent intentions will have a very low confidence rating less than 30% and owner will be alerted.
Shows a high confidence rating of 72% as the face is recognised.
As soon as the sytem detects an unrecognised face, it generates an alert and sends an email from safeHome account to the owner of the house along with the image of the stranger. Thus, the owner can take immediate action if any criminal or unknown person breaks into the house without their knowledge.
Monitors and ensures complete safety of home when the members are not at home. If it detects an unrecognised person, it displays a very low rating and mails the owner to warn them.
Unrecognised:
Recognised:
Monitors and ensures complete safety of home when the members are at home.
In accordance with the principles of Agile Methodology, this project has been built in four sprints spread evenly over the four weeks of the mentorship program. I had a roadmap ready for each week beforehand so as to systematically build the project in a planned manner.