This file contains Project for Face Detection Analysis using Machine Learning Algorithms. # Dataset used for this project is MIT Center For Biological and Computational Learning (CBCL) lab.
• Training set contains 2,429 faces, and 4,548 non-face images • Test set contaings 472 faces, and 23,573 non-face images • Each image file is 19 × 19 grayscale image, and is stored as a pgm file.
That means – – Each image sample is 19 pixel wide and 19 pixel high. – So, each image sample is a 361 dimensional vector, where each ele- ment in the vector contains the grayscale value of a pixel from the image. – A grayscale value of a pixel takes any (discrete integer) value from the range {0,1,···,255}
Implemented the Face Detection using Naive Bayes Classifier, Batch gradient descent based logistic regression (BGD-LR), Stochastic gradient descent based logistic regression (SGD-LR).