README.md

Gesture-Controller

Team Member:-

• Meet Variya - https://github.com/meetvariya

Problems:-

• Many of the Disabled People find it difficult to write using keyboard ,Also in Covid Situation many people don’t want to touch stuff that’s related to Covid.
• Sometimes when Mouse or Keyboard doesnt Work Properly Incase of a laptop is a Difficulty.

Solution (Overview of project):-

• We make a Gesture Based Keyboard Interface using Webcam so that one can type things in a new Interactive way. The Way it Works is using your Hand gesture and Eye tracking to use the Gui keyboard and mouse to Interact With it.

Key Features:-

• Developed python exe file setup using tkinter to fully control mouse & keyboard movements by human hand & eye gestures.
• It works with the help of different libraries and packages like OpenCV, Mediapipe, Cvzone, Dlib, Haar Cascade, etc.
• Also included Voice & basic storage features. Helpful in such a circumstances like COVID-19.

Working:-

Model Detects Face and Hand points using following modules:

dlib's 68 points facial detection:

• The mouth can be accessed through points [48, 68].
• The right eyebrow through points [17, 22].
• The left eyebrow through points [22, 27].
• The right eye using [36, 42].
• The left eye with [42, 48].
• The nose using [27, 35].
• And the jaw via [0, 17].

HandtrackingModule's 21 points in hand:

• 4,8,12,16,20 - top of all five fingers, thumb top to pinky fingure top respectively.

Different packages:

cvzone's HandTrackingModule to detect hands. Dlib's shape predictor(shape_predictor_68_face_landmarks.dat) and haarcascade classifier(frontal_face_detcetor.xml) to detects face imutils to fetch eyes landmarks from dlib's 68 points facial detection. Eye-aspect-ratio(EAR) for detect blinking. pynput for controlling keyboard events. pyautogui for mouse events

Dependencies:-

• CvZone
• Pynput
• HaarCascade
• PyAutoGui
• Tkinter
• Mediapipe
• auto-py-to-exe
• Dlib

References:-

• Demo Link

• Reference link for .dat file

Gesture.Keyboard.mp4