https://github.com/elvis-not-presley-one/animal-image-recognition-model
An Image Recognition Model That uses a Convolution Neural Network with Tensorflow to be able to tell an animal from a non-animal image
https://github.com/elvis-not-presley-one/animal-image-recognition-model
cnn convolution-neural-network data-science dataset image-classification image-recognition machine-learning neural-network python tensorflow
Last synced: 3 months ago
JSON representation
An Image Recognition Model That uses a Convolution Neural Network with Tensorflow to be able to tell an animal from a non-animal image
- Host: GitHub
- URL: https://github.com/elvis-not-presley-one/animal-image-recognition-model
- Owner: Elvis-Not-Presley-One
- License: mit
- Created: 2024-12-04T22:29:35.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-01-16T23:22:50.000Z (over 1 year ago)
- Last Synced: 2025-04-03T15:12:54.203Z (over 1 year ago)
- Topics: cnn, convolution-neural-network, data-science, dataset, image-classification, image-recognition, machine-learning, neural-network, python, tensorflow
- Language: Jupyter Notebook
- Homepage:
- Size: 7.3 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Animal-Image-Recognition-Model
This is a image recognition model project that uses Tensorflows CNN(Convolutional Neural Network) architecture.
- This was a class project, code was writen by four people
# what is a Convolutional Neural Network(s)
- A Convolutional Neural Network is a type of neural network architecture that is popular use cases are for:
- image reconition
- object detection
- image classification
- video recognition
- A Neural Network is composed of 3 Layers :
- Input Layer: Takes in raw data, like pictures or numbers.
- Hidden Layers: Process and combine the data to find patterns.
- Output Layer: Gives the final result, like “This is a cat.”
- Learning Through Mistakes:
- At first, guesses are random. The network learns by fixing mistakes, repeating until it gets better.
- Connection Strength:
- The links between neurons are like wires. Learning strengthens important connections and weakens less useful ones.
# How Tensorflow Works
- For this project we decided to use tensorflow as our main lib
- TensorFlow helps create and train models for applications like image recognition, language processing, and predictions.
- TensorFlow uses "tensors," which are multi-dimensional arrays, to process and analyze data. It offers user-friendly tools for beginners and advanced features for experts and works on different platforms, including computers and mobile devices, making it flexible for various projects.
# Results
- The Model Produced around a ~98% accuracy
# Dataset
- Since github will not allow me to include the dataset we made the link for the Kaggle page will be below
- About 3gb worth of images:
- comprising of Googles recaptcha v2 images for most non-animal images and landscapes;
- The other half comes from 4 diffrent large animla datasets with over 50+ diffrent species
https://www.kaggle.com/datasets/tylerelvis/animal-vs-non-animal-image-recognition-dataset
# Rrequirements
- All Libs that need to be downloaded to the latest verstion
```!pip install tensorflow opencv-python matplotlib```
# Links That Helped Us Along The Way
Important links:
- https://www.datacamp.com/tutorial/cnn-tensorflow-python
- In-depth info about tensors and how image recognition works
- https://www.youtube.com/watch?v=jztwpsIzEGc
- Great video, guy goes really in-depth with everything
- https://www.tensorflow.org/api_docs/python/tf/all_symbols
- TensorFlow Documentation
- https://github.com/billy-enrizky/TensorFlow-Image-Classification/blob/main/index.ipynb
- Really good example project