GestureTrack

Extracting Data from the input

Basic Input

Basic frames captured from the webcam

Image Transformtion

The image resolution is huge in size and takes considerable amount of processing power for simple tasks. Hence we apply Haar wavelet transform

Applying filters to the given Image

  1.     Blur Image
  2.               
  3. Threshold Image
  4.  
  5.                     Drawing Contours
  6.  
  7.                  Find convex hull and convexity defects
  8.                
  9.  

Recognizing Movements

The underlying algorithm for the used motion detection

  • First two frames of the video signal are compared. If the viewer moves his/her head the frames show small differences. 
  • Then the average position of all the changed pixels (blue pixels) will be computed (small red cross). This moving average shows the average position of all movements in the screen – usually the head of the user. 
  • source code available at github.com/namitjuneja

Source Coding the data 

and decoding the data at arduino 

  • We have an [NxN] pixel array 
  • Each element has an intensity value between 0 - 255

 

  • We convert the image into an [N^2 x 1]   basis vector to confine the signals within a given range

 

 

 

 

  • The wave form coding can be generated using various image parameters such as individual pixel intensity

 

  • But in this particular case difference in the pixel intensity of two adjacent pixels was chosen in order to reduce the processing time.

 

 

Quantization of the signal

  • The resolution of the intensity heat map is reduced

 

  • The range of intensities are divided according to the schematic shown and are bit by bit serially to the arduino board. 
0 - 50    : 00

50 - 100  : 01

150 - 200 : 10

200 - 150 : 11

Serial Transfer

  • The transfer of the generated intensities was transfered to the arduino using the pySerial library , 2bits at a time

 

Output at the Sensors

  • The transmitted code was decoded using the same code book as mentioned in the earlier slide

 

  • Subsequently a analog signal was generated which was transmitted to the LEDs and other sensors using Pulse Width Modulation technique

 

  • Hence corresponding to the PWM signal which varies according to the direction of movement, the intensity of light varies accordingly

 

 

Made with Slides.com