Krishna Chaitanya Gurijala
Computer Science Department
PhD Student
SunySB# 106887301

Introduction to Computer Vision
Assignment# 1
Problem# 1
View the solution to the first problem by downloading the following .docx file:  Problem 1

Problem# 2
To view the .avi files below, you can download and install the vlc player from here (Thank you):  VLC player

Input test sample video given:  input sample video
Number of frames in the test sample input video given = 592


Codes : final.m   gaussdensity.m
Output video generating only skin layer(Due to memory constraints, only alternate frames were considered -- has only 296 frames):  output video
Output video generating only skin -- all containing all frames:  Output with all frames
Output video generating only skin -- black and white mode:  Output in b & w
Output video generating only skin in real mode(texture):  Output in real mode

Concept
Using the Gaussian Distribution, we calculate the probability matrix, which can be interpreted as "posteriori probabilities": P(skin|RGB). Finally we generate the output video of only the skin layer.

Reading References
"Statistical color models with application to skin detection",  Michael J. Jones, James Rehg, International Journal of Computer Vision, 1999.
"A Survey on Pixel-Based Skin Color Detection Techniques",  Vladimir Vezhnevets, Vassili Sazonov, Alla Andreeva,  Proc. Graphicon-2003, 2003.