Update Readme

ivclab · Oct 23, 2018 · 0c9f6c9 · 0c9f6c9
1 parent ef3e5e9
commit 0c9f6c9
Show file tree

Hide file tree

Showing 14 changed files with 43 additions and 51 deletions.
diff --git a/C-GLO/README.md b/C-GLO/README.md
diff --git a/Detection/README.md b/Detection/README.md
diff --git a/README.md b/README.md
@@ -3,21 +3,55 @@ Official Implementation of [Stingray Detection of Aerial Images Using Augmented
 
 Created by [Yi-Min Chou](https://github.com/yyyjoe), Chien-Hung Chen, Keng-Hao Liu, Chu-Song Chen
 
-## Usage
-    ConditionalGLO
-            ├──────────── C-GLO
-            └──────────── Detection
+## The overview of data augmentation
 
+![alt tag](./assets/overview.png)
 
-1.Clone the Mixed-Bg-Fg-Synthesis repository:
+1. Crop the object patches (rotate and flip to augment data) and randomly crop the background patches to establish the dataset for training conditional GLO.
+2. Training conditional GLO and use the well-trained model to generate the fake stingray image.
+3. Paste the generated stingray patches to original positions.
+4. Use the augmented data generated from C-GLO to train detection models. 
+
+## Conditional-Generative-Latent-Optimization
 
-    $ git clone --recursive https://github.com/ivclab/ConditionalGLO.git
+![alt tag](./assets/cglo-1.png)
 
+## Prerequisition
+- Python 2.7
+- [TensorFlow 1.4.0 or higher](https://github.com/tensorflow/tensorflow)
 
-2.Follow the instruction in [C-GLO](https://github.com/ivclab/Mixed-Bg-Fg-Synthesis/tree/master/C-GLO) and get the augmented training data.
-
+## How to Run
 
-3.Use the augmented data generated from C-GLO and train on Stingray [Detection](https://github.com/ivclab/Mixed-Bg-Fg-Synthesis/tree/master/Detection).
+- Clone the Mixed-Bg-Fg-Synthesis repository:
+```bash
+$ git clone --recursive https://github.com/ivclab/ConditionalGLO.git
+```
+
+- Install required packages:
+```bash
+$ pip install -r requirements.txt
+```
+
+- Download Stingray Data:
+```bash
+$ python download.py
+```
+
+- Run the training code:
+```bash
+# The training result will be saved in `./logs/FOLDER_NAME/`
+$ python main.py --is_train=True
+```
+
+- Run the testing code:
+```bash
+# The testing result will be saved in `./logs/FOLDER_NAME_test/`
+$ python main.py --is_train=False --load_path=FOLDER_NAME
+```
+
+## Experimental Results
+Original background image(top), mixed background and foreground synthesis generated by C-GLO (bottom)
+![alt tag](./assets/Fig5-1.png)
 
 
 ## Citation

diff --git a/C-GLO/assets/Fig5-1.png → assets/Fig5-1.png b/C-GLO/assets/Fig5-1.png → assets/Fig5-1.png
diff --git a/C-GLO/assets/cglo-1.png → assets/cglo-1.png b/C-GLO/assets/cglo-1.png → assets/cglo-1.png
diff --git a/Detection/assets/overview.png → assets/overview.png b/Detection/assets/overview.png → assets/overview.png
diff --git a/C-GLO/config.py → config.py b/C-GLO/config.py → config.py
diff --git a/C-GLO/data_loader.py → data_loader.py b/C-GLO/data_loader.py → data_loader.py
diff --git a/C-GLO/download.py → download.py b/C-GLO/download.py → download.py
diff --git a/C-GLO/main.py → main.py b/C-GLO/main.py → main.py
diff --git a/C-GLO/models.py → models.py b/C-GLO/models.py → models.py
diff --git a/C-GLO/requirements.txt → requirements.txt b/C-GLO/requirements.txt → requirements.txt
diff --git a/C-GLO/trainer.py → trainer.py b/C-GLO/trainer.py → trainer.py
diff --git a/C-GLO/utils.py → utils.py b/C-GLO/utils.py → utils.py