Portfolio Project : Smartfeel IOS APP

We started our educational career with Holberton School Tunisia 2 years ago . In the first year, we studied algorithm techniques from 0 to hero, from Low level programming to high level programming and Devops knwoledge , we accomplished the first level with a software engineer degree.

After that , I chose my specialty. It’s always been my dream to be a machine learning engineer .

We learned Supervised learning , unsupervised learning and reinforcement learning using many tools, libraries , frameworks and python native to create neurons deep learning algorithms and data visualizations. Tensorflow keras and OpenCV are helpful tools .

Finally, I reached the last Steps to be a machine learning engineer .Today, I’m working on my Portfolio project for my graduation. My project with SMARTFEEL, it’s a mobile application used in the background of any smartphone Android or iOS to detect and scan faces of users when utilising any app. After that, the application collects data and communicates with the model of deep learning to recognize the user’s facial expressions and detect emotions.

And in order for me to be able to do that I learnt new technologies like REACT NATIVE to develop the app. I integrated the models to detect faces with OPENCV and created deep learning models with training and testing using Keras and tensorflow .

To complete this project I followed 5 steps :

  • Scraping Data using Beautifulsoup4: with beautiful and Python we collect data images from google for any data we want. beautifulSoup helps us take divs <div></div> and balises in html from any link we want using request .
  • Cleaning and sorting the data by using Data Analyst techniques
  • Creating a deep learning model , training and testing it using TensorFlow keras openCV to detect(facial expressions and emotions.
  • Creating the mobile application and integrating it with the model of deep learning .

Finally i want to share with all who want to learn and testing this projects this is the link of repo in Github :