Real-time Object Detection using YOLO

Real-time Object Detection using YOLO algorithm. You only look once (YOLO) is a state-of-the-art, real-time object detection system that is extremely fast and accurate. You can read more about it here.

Getting Started

In order to make detections, you will need to install the required python packages. I recommend using virtualenv for this one if you're into that stuff:

pip install -r requirements.txt

This project uses YAD2K helper to build the Darknet19 model using YOLO pre-trained weights and configuration which you will need to download in order to make detections on images:

git clone https://github.com/allanzelener/YAD2K.git
cd YAD2K
wget http://pjreddie.com/media/files/yolo.weights
wget https://raw.githubusercontent.com/pjreddie/darknet/master/cfg/yolo.cfg

Now, you can go ahead and build the Darknet19 model using:

python yad2k.py  yolo.cfg yolo.weights yolo_model.h5 --fully_convolutional

Once you have the pre-trained model yolo_model.h5, move it over to model_data/ dir in your project and start detecting all kinds of stuff in real-time:

python main.py --realtime

You can also run the detection on images, all you need to do is provide the path of the image (results in images/out/):

python main.py --image-path images/test/cars.jpg

And you're done.

Build With

• Python
• Keras
• TensorFlow
• OpenCV

References

• You Only Look Once: Unified, Real-Time Object Detection
• YOLO9000: Better, Faster, Stronger

Credits

• YOLO: Real-Time Object Detection
• YAD2K