https://github.com/sibyjackgrove/cnn_for_bird_classification

CNN to classify images of birds into Cardinal and Indigo
https://github.com/sibyjackgrove/cnn_for_bird_classification

classify-images cnn tensorflow tensorflow-experiments

Last synced: 5 days ago
JSON representation

CNN to classify images of birds into Cardinal and Indigo

• Host: GitHub
• URL: https://github.com/sibyjackgrove/cnn_for_bird_classification
• Owner: sibyjackgrove
• License: gpl-3.0
• Created: 2017-02-21T21:55:13.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-02-21T22:29:49.000Z (almost 8 years ago)
• Last Synced: 2024-11-09T06:38:22.579Z (2 months ago)
• Topics: classify-images, cnn, tensorflow, tensorflow-experiments
• Language: Jupyter Notebook
• Size: 10.2 MB
• Stars: 0
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# CNN_for_classifying_bird_images

Convolutional Neural Network to classify images of birds into Cardinal and Indigo. Only TensorFlow operations are used to read image files, and build CNN.

# Dependencies

TensorFlow 1.0, Numpy

# Operations

tf.layers, tf.image, tf.gfile, tf.losses

# CNN Architecture

2 Convolutional layers with ReLU activation, 2 Max Pooling Layers, 1 Fully Connected Layer with Softmax Activation

# Training

Training data: 72 images,

Validation data: 24 images,

Testing data: 4 images

# Classification - 3 Classes

0: Cardinal; 1: Indigo; 2: None of these

# Dataset

Source: CUB-200 available at http://www.vision.caltech.edu/visipedia/CUB-200.html