minzhou1003/deep-learning-miniproject

This is the mini project using deep learning model to recognize roses and sunflowers.

deep-learning-mini project

This is the mini project from EC601.

Author: Min Zhou

Email: [email protected]

Introduction

In this mini project, I trained two different deep learning models to recognize roses and sunflowers.

1. A simple model(without conv layer): only contains one simple flatten layer and two densely-connected NN layers.
2. VGG-16 model: using the VGG-16 structure.
3. Training, validation and test dataset:

Number of training examples = 1092
Number of validation examples = 274
Number of test examples = 152
X_train shape: (1092, 64, 64, 3)
y_train shape: (1092,)
X_val shape: (274, 64, 64, 3)
y_val shape: (274,)

File Instruction

• rose folder: images of rose.
• sunflower folder: images of sunflower.
• test_images folder: store all the images you want to test using the trained model. There are three images already if you don't have test images.
• flower_recognition_using_deep_learning.ipynb: the Jupyter Notebook for developing and training the model.
• simple_flower_recognition.py: the python API to run the trained "simple model".
• simple_model.h5: trained model weights.
• simple_model.json: trained model.
• vgg16_model.h5: trained VGG-16 weights(This file is too big can't be uploaded to github)
• vgg16_model.json: trained VGG-16 model.
• requirements.txt: necessary python libraries.

Compare models:

The test accuracy of the simple model and VGG-16 model are 85.5% and 55.9% respectively. There are some main facts caused the lower accuracy of VGG-16 model:

1. Overfitting
   • Lack of data.
   • The model is too complex for this simple binary classification problem.
2. Optimal hyperparameters need to be tuned in the future.

Installation:

1. Download this repository:

git clone https://github.com/minzhou1003/deep-learning-miniproject.git

2. Put your test image to the test_images/ folder

There are three images already if you don't have test images.

3. Set up and activate virtualenv inside that folder.

cd deep-learning-miniproject
virtualenv --python python3 env
source env/bin/activate

4. Install python libraries:

pip install -r requirements.txt

5. Run the python API of the simple model:

python simple_flower_recognition.py

6. Sample output:

$ python simple_flower_recognition.py
Using TensorFlow backend.

Did you put all test images to the "/test_images" folder? ("yes"/"no") yes
2018-10-22 01:33:58.746880: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA

Model successfully loaded from disk!

Image: 15951588433_c0713cbfc6_m.jpg, Predict result: rose

Image: 15965652160_de91389965_m.jpg, Predict result: rose

Image: 15972975956_9a770ca9dd_n.jpg, Predict result: sunflower

Goals:

1. Use TensorFlow (or any tool you prefer) to recognize between two classes of objects.

• Roses and Sunflowers

2. You need to capture the images yourself and tag them. You can use any Tagging system (e.g, https://hubs.ly/H0dktLv0 Neurala).
3. You need to design your own Training, Testing, and Verification sets.
4. You need to compare between two different systems based on the literature review you did and/or reading the material provided Bonus: [x] Provide an API with example code for another developer to use your system