This is a simple Flask app that uses a convolutional neural network(CNN) model to classify images as either a dog or a cat.
The model is built using the MobileNet architecture and trained on the Dogs vs. Cats dataset provided by Microsoft Research, which contains 25,000 images of dogs and cats labeled as 1 (dog) or 0 (cat).
HOME PAGE(index.html)
PREDICT BUTTON
PREDICTION PAGE(predict.html)
The main goals of this project are to:
- Build a deep neural network using MobileNet to classify images of dogs and cats
- Create a Flask application that allows users to upload photos and receive real-time predictions of whether the photo contains a dog or a cat
- Tensorflow
- Keras
- MobileNet
- h5py
- Flask
- Bootstrap
- Heroku
The final model achieved an accuracy of 98% on the test set, ** which is good performance for the dataset we got as 2000. The model is saved in the cats_dogs_classifierv2.h5 file, and can be loaded and used for future predictions.
A CNN is a deep learning model that is commonly used for image classification tasks. In this project, we used the MobileNet architecture, which is optimized for mobile devices and has a small memory footprint. The model is trained on the Dogs vs. Cats dataset using TensorFlow and Keras. The MobileNet of the model is available here. Also Inception of the model is available here.
You can find details in the article.
docker push badl7/catdogflaskapp
Here are some ideas for further development:
- Deploy the app using Docker
- Add more models, such as
- VGG-16
- ResNet50
