Simple binary image classifier is a Machine-Learning approch with the purpose of classification is either be a yes/no decision so-called binary classification where a given item is predicted as class 0 or class 1, or several classes referring to multi-class classification.
- Raw data collection
- data-preprocessing and segmantaion
- Model preprocessing
- Training
- Prediction
In this project I implemented simple binary classification over 200 image data-sets collected from google images.
The training includes minimization of cost function by continuous iterations over datasets while performing Gradient decent in each step and minimizaing parameters to fit the training sets.
The cost function over 100 iterations
