TensorFlow Binary Image Classification using CNN's

This is a binary image classification project using Convolutional Neural Networks and TensorFlow API (no Keras) on Python 3. Read all story in Turkish.

It is a ready-to-run code.

Dependencies

pip3 install -r requirements.txt

Notebook

jupyter lab Binary_classification.ipynb or jupyter notebook Binary_classification.ipynb

Data

No MNIST or CIFAR-10.

This is a repository containing datasets of 5000 training images and 1243 testing images.No problematic image.

train_data_bi.npy is containing 5000 training photos with labels.

test_data_bi.npy is containing 1243 testing photos with labels.

Classes are table & glass. Classes are equal.

Download pure data from here. Warning 1.4 GB.

Training

Training on GPU:

python3 binary_image_classification_GPU.py

Training on CPU:

python3 binary_image_classification_CPU.py

Architecture

AlexNet is used as architecture. 5 convolution layers and 3 Fully Connected Layers with 0.5 Dropout Ratio. 60 million Parameters.

Results

Trained 5 epochs. Accuracy, AUC and Loss graphs are below:

Predictions