asmedeus998/yolo_v3_object_detection

Project Deep Learning

Yolo v3 Object Detection in Tensorflow

Yolo v3 is an algorithm that uses deep convolutional neural networks to detect objects.

Getting started

Prerequisites

This project is written in Python 3.6.6 using Tensorflow (deep learning), NumPy (numerical computing), Pillow (image processing), OpenCV (computer vision) and seaborn (visualization) packages. I will be using anaconda to install all the library.

conda env create -f enviroment.yaml

Downloading official pretrained weights

Let's download official weights pretrained on COCO dataset.

!wget -P weights https://pjreddie.com/media/files/yolov3.weights

Save the weights in Tensorflow format

Save the weights using load_weights.py script.

python load_weights.py

Running the model

Now you can run the model using detect.py script. Don't forget to set the IoU (Intersection over Union) and confidence thresholds.

Usage

python detect.py <images/video> <iou threshold> <confidence threshold> <filenames>

Images example

Let's run an example using sample images.

python detect.py images 0.5 0.5 data/images/bali_street.jpg

Then you can find the detections in the detections folder.

python detect.py video 0.5 0.5 data/video/Street.mp4

The detections will be saved as detections.mp4 file.

To-Do List

• Model training in GPU

Acknowledgments

• Yolo v3 official paper
• What’s new in YOLO v3?
• A Tensorflow Slim implementation
• ResNet official implementation
• DeviceHive video analysis repo
• Evening Walk ~ Thamrin Nine to Sudirman Plaza thru Setiabudi Astra MRT Station ~ Jakarta