A modular pipeline to extract several facial features from videos such as face landmarks, eye gaze direction, head pose and Action Units

Scripts for importing pupil labs eye-tracking data into matlab and extracting pupil, gaze data etc.

purpose - extract face -> extract eye -> extract pupil( or with iris) -> extract gaze for eye tracking

A CNN-based drowsiness detector with Dlib facial feature extractor. Designed a sub-network based on lightweight architectures for yawn detection. Implemented a CNN based on attention mechanism for eye gaze estimation

Saliency system to model the behaviour of human eye gaze using deep learning. The model uses pre-trained VGG like backbone network and upsamples the high level features extracted from various intermediate layers using bilinear interpolation.

No description available

This repository contains a lightweight pipeline for extracting and analyzing eye-tracking data from gameplay recordings overlays. Instead of relying on proprietary eye-tracking software, the system uses computer vision to detect the eye gaze indicators in recorded videos and converts it into structured gaze data that can be analyzed with Python.

GazeInsight is a specialized data analysis pipeline designed to extract meaningful insights from facial data, with focus on eye-gaze behavior from WheelSimPhysio-2023 dataset

In this project, I have captured video streaming of user through the webcam of computer, using python library packages ( dlib and open-cv) extracted the facial information and detected several eye gaze information. Finally, the eye gaze information were utilized for an on-screen keyboard and three-command based operation.

EEG labelling pipeline using eye-tracking data to automate annotation. Extracts fixations (I-DT), clusters attention with DBSCAN, and models gaze using GMMs. Generates heatmaps and maps gaze to EEG channels/time, producing structured labels for machine learning and analysis.

Python notebook for analyzing Tobii eye-tracking data and generating gaze heatmaps over visual stimuli (e.g., faces). The pipeline synchronizes gaze data with XDF/LSL experiment markers, extracts stimulus presentation windows, and visualizes fixation patterns directly on stimulus images.

Pupil (Eye) Detection with Mediapipe and OpenCV: Real-time eye detection and cropping using Mediapipe Face Mesh with iris tracking and OpenCV. Extracts and displays left and right eyes from webcam video with dynamic bounding boxes. Useful for eye-tracking, gaze estimation, and attentive user interface projects.

A password system which works on image processing and extracting eye images from the facial image. Different eye gaze positions are fed into a trained convolutional neural network to check if a image selected is same as password image selected at the time of Registration.

This project is an automated interview video analysis platform built using Streamlit, OpenCV, MediaPipe, and Transformers. It analyzes uploaded video files to: 1. Extract audio and perform speech-to-text conversion. 2. Conduct sentiment analysis on the extracted text. 3. Detect eye gaze and potential cheating.

This repository contains Python scripts for preprocessing, analyzing, and visualizing eye-tracking data. It includes tools for extracting temporal windows, detecting fixations and saccades, computing dispersion and speed metrics, generating heatmaps, and performing statistical tests (t-test, Wilcoxon, Shapiro-Wilk) on gaze and behavioral data.