Saravanan S's repositories
EyeTracking-to-find-the-AreaofInterest
This project is about trasking the user eye movement and finding the particular Area of Interest from the whole screen. Its application starts from using it give more personalised advertisements based on which region our eye is looking upon, It can be used to detect our concentration level by monitoring our eye movement etc
mastering_ros
This repository contains exercise files of the book "Mastering ROS for Robotics Programming"
TextDetection-using-CNN-from-scratch
The code detects text by creating a Convolution Neural Network to Classify digits from 0 to 9. The Training code is written from scratch and it trains about 10000 images of 10 different classes . Testing script can deduce the trained model data to use along with a webcam to detect Digits from 0 to 9.
Diabetic-Retinopathy
Diabetic Retinopathy(DR) is a leading cause of blindness, It's caused by damage to the blood vessels of the light-sensitive tissue at the back of the eye (retina). a model is developed using the pretrained 'Efficinet-b3' architecture. A custom Neural network present on the top of efficinet model will help in using both the image of eye taken during the examination so that the accuracy improves. Various image data generation techniques are used in order to cope with the class imbalance problem.
Emotion-recognition-in-conversations
This repo contains my work on emotion recognition and classification in a conversation using multimodal Deeplearning.
Home-Automation-using-ESP8266-and-Alexa
Alexa triggers the Nodemcu through http protocol which in turn triggers the Ir transmitter to send specific signals to control all the Ir appliances in the room. nodered platform has been established which runs on ec2 instance(AWS) which gives stastistical data on units of current consumed etc. Provision is made to control your appliance through internet from anywhere in the world
Machine-Learning-Collection
A resource for learning about ML, DL, PyTorch and TensorFlow. Feedback always appreciated :)
MELD
MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversation
open-cv
This repository has bunch of open-cv files which comes in handy while writing a program
Sudoko-Solver
This is an open-cv code which uses a Convolution Neural Network to classify digits from 0 to 9, using keras for training of the model. It was written from scratch, and is available under the name "myModel". This code takes an unsolved sudoku image as input, processes the image detect the numbers, solves the sudoko puzzle and shows us output at each stages.
WebTechnology_Assignment1
This repo contains the Assignment 1 for the web technologies course work