CodeHers

Inspiration

We embarked on our initial learning trip into the field of data science and deep learning after being motivated by the All India Women ONLY Hackathon'23 for Tech For Good. To take our first peek under the hood of Data Science and see what we could build!

What it does

It provides a seamless user interface experience, allowing user to authenticate themselves for transactions or OTP verification by simply using their facial features, eliminating the need for passwords or PINS.

User initiate transaction by application which captures the user's image through the camera if face is not detected the application again goes for image capturing.If face is detected facial features extraction is done by OpenCV and TensorFlow, Dlib. It checks if against the enrolled data in the database if necessary. If the image doesn't matches with the database unrecognised face alert is generated. And if it matches system verifies the user and approves for the OTP request, a real time transaction monitoring and anomaly detection mechanisms during OTP verification is done i.e., if fraud is detected it triggers an alert if not then transaction is done successfully.

How we built it

We built face recognisation by the use of OpenCV which is used for study of deep learning and we did this in Google colab which will compare it to positive and negative biometric. We collected data in .jpg image our model is train to verify face of user and deal it to forward process MATLAB ,numpy deep learning, google colab

Challenges we ran into

We ran into many challenges as we simultaneously worked to understand Deep Learning and deploy our project as we initially intended. We also struggled while training and face verification. We did finally figure that out but our UI/UX suffered due to time constraints. We had limitation of Google colab as it was hard for webcam to access plt.imageshow.

Accomplishments that we're proud of

We're proud of everything we learned during this short duration time, even the aspects of our application that we struggled with and those we couldn't get to work. We really stretched ourselves into unchartered, for us, the AI landscape!

Even though our Facial transaction application doesn't currently work, we are proud of the working components as well as our determination to learn and build an AI/ML app.

What we learned

We learned that there's a lot to learn about Data science! We're looking very forward to exploring the possibilities within the Data science Infrastructure. We learned how to work in a diverse team working remotely and gaining experience and so much memory.

What's next for CodeHers

We're going to do some more reading and research about OpenCV with the intent to bring CodeHers into a fully functional deep learning application where we improves the accuracy and speed. Ensuring data privacy , in near future we are planning for broader picture by integrating other biometric authentication methods like iris and retina detection. As of now we made a dummy transaction where only one can transact in near future we would extend the transaction for multiple user.