Facial-Keypoint-Detection-
Solution to kaggle competition (https://www.kaggle.com/c/facial-keypoints-detection ) using Python 3 and Pytorch Library
Data Description
Data includes 1 * 98 * 98 images and there are 30 target lables we need to predict
Model Architecture
-
Multilayer perceptron model
96 * 96 ==> 128 ==> 64 ==> 30
-
CNN Model 1 * 96 * 96 ==> 16 * 96 * 96 ==> 16 * 48 * 48 ===> 32 * 48 * 48 ===>32 * 24 * 2 4 ===>64 * 12 * 12 ==>64 * 12 * 12 ===>1024 ===>30