Udacity Capstone project in the Data Scientist Nanodegree! In this project, Explain how to build a pipeline that can be used within a web or mobile app to process real-world, user-supplied images. Given an image of a dog, your algorithm will identify an estimate of the canine’s breed. If supplied an image of a human, the code will identify the resembling dog breed.
Along with exploring state-of-the-art CNN models for classification Each model has its strengths and weaknesses, and engineering a real-world application often involves solving many problems without a perfect answer.
- Step 0: Import Datasets
- Step 1: Detect Humans
- Step 2: Detect Dogs
- Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
- Step 4: Use a CNN to Classify Dog Breeds (using Transfer Learning)
- Step 5: Create a CNN to Classify Dog Breeds (using Transfer Learning)
- Step 6: Write your Algorithm
- Step 7: Test Your Algorithm
- Clone the repository and navigate to the downloaded folder.
git clone https://github.com/udacity/dog-project.git
cd dog-project
-
Download the dog dataset
-
Download the human dataset.
-
Donwload the VGG-16 bottleneck features
-
Libraries
Keras OpenCV Matplotlib Numpy Tensorflow tqdm -
File descriptions
dog_app.ipynb: Jupyter notebook containing the main algorithm, findings and results. dog_app.html: A copy of dog_app.ipynb in html format. Haarcascades folder: contain the opencv haarcascade XML file for face detection. images: few images for the testing prediction algorithm. -
Findings
- The model achieved a test accuracy of 83.4928%.
- Data is not augmented some of the breeds have similar colors and shapes but differ in size, so some of the pic where unclassified
- The model could use some improvements on its ability to classify pictures with data augmentation and using the other trained model
-
Open the notebook.
jupyter notebook dog_app.ipynb
Detailed Discussion and Result of the project are in the article in medium Dog Breed classifier