π¦Before that I was working as an ML Engineer at @IceCream Labs.
β³ Wanna know about my internships, I have interned as SWE intern and Data Scientist at @OpenMainFrame and @IceCream Labs respectively.
π¬ Now, after all that corporate talking, I am a guy who likes to kill time by doing something fun, and programming is one of them. If you wanna know more about me, connect with me on social media platforms π
A webapp that takes patients ECG readings from their wearable devices and uses a LSTM-based-Autoencoder to detect if there is an anomaly in their heart readings. If the data is abnormal, our tool will send the important data to the patient's electronic medical records and provide them with a simple βnormalβ vs βabnormalβ result.
I have build a facial emotion detector web app. In this project, I used keras API with tensorflow as backend, OpenCV and flask. This project has 3 distinctive parts:- EXPRESSION REGRESSOR #Using keras I trained a model with images of 7 different facial expressions (happy, angry, sad, disgust, neutral, surprise, fear). During this model development, I faced class imbalance problem which I tackled with ImageDataGenerator. FACE DETECTOR #The next part was capturing images from my webcam and detecting faces. I accomplished this task by OpenCV's face detector (haarcascade_frontalface_default.xml). So, basically the faces that were detected were feeded to my expression regressor model and the predicted emotion were shown on the camera screen. MODEL DEPLOYMENT #The final part was model deployment, for which I used flask framework. An interesting thing about this project is that it also detect facial emotions in videos or images with just one line change in code.