Happy-Sad-Face-Classifier
- 0 Collaborators
CNN model to classify happy and sad faces. ...learn more
Project status: Concept
Overview / Usage
Implementation of a CNN model in Keras to identify happy and sad faces.
Methodology / Approach
The model in a nutshell :
INPUT
-> | zero_padding((3, 3)) -> conv2D(32, (7, 7)) -> batchNorm -> relu -> maxPool((2, 2)) |
-> | zero_padding((2, 2)) -> conv2D(64, (5, 5)) -> batchNorm -> relu -> maxPool((2, 2)) |
-> | zero_padding((1, 1)) -> conv2D(128, (3, 3)) -> batchNorm -> relu -> maxPool((2, 2)) |
-> | conv2D(64, (1, 1)) -> batchNorm -> relu |
-> | zero_padding((1, 1)) -> conv2D(32, (3, 3)) -> batchNorm -> relu -> maxPool((2, 2)) |
-> | Flatten -> dense_layer_1(128, sigmoid) -> dense_layer_2(32, sigmoid) -> dense_layer_3(1, sigmoid) |
-> OUTPUT
Looks messy here. The one in the GitHub repo readme file might look better.
Technologies Used
Keras, NumPy