jangocheng / open_cv_ui2code

open_cv_ui2code

Steps To Detect Component

1. Install and setup tensorflow [https://www.tensorflow.org/install/]
2. Activate Tensorflow Environment
3. Install following packages

• imutils
• numpy
• opencv-python
• matplotlib
• pandas
• pillow

To install
pip install pandas

4. Set Environment Variables
   IMAGE_SIZE=224
ARCHITECTURE="mobilenet_0.50_${IMAGE_SIZE}"

5. To Train
   python -m scripts.retrain \ --bottleneck_dir=tf_files/bottlenecks \ --how_many_training_steps=500 \ --model_dir=tf_files/models/ \ --summaries_dir=tf_files/training_summaries/"${ARCHITECTURE}" \ --output_graph=tf_files/retrained_graph.pb \ --output_labels=tf_files/retrained_labels.txt \ --architecture="${ARCHITECTURE}" \ --image_dir=images/

6. To Predict Component

python detect_component.py --image=predict/web.png

To Render on Webpage

1. cd to myapp
2. npm install
3. npm start
4. Now detect components for a image and it will automatically reflect on Webpage