gknova61/Object-Detection-In-Media

Computer Vision/Detecting Objects in Videos using YOLO

Object Detection In Media Using YOLOv3

CPSC 393 Machine Learning Final Project

Carlos Amesquita, Everett Yee, Keanu Kauhi-Correia

In order for YOLOv3 to work, the weights file must be put into yolo-coco folder. It can be found here

Instructions for Replication

Directory Structure should look like this:

root

├── annotations

├── images

├── output

├── yolo-coco

││││ ├── coco.names

││││ ├── yolov3.cfg

││││ └── yolov3.weights

├── accuracyTracker.py

└── yolo.py

Configurations

There are two files that have a configuration.

• yolo.py (lines 18-25)
• accuracyTracker.py (lines 14-18)

1. If you have annotations for truth comparisons, put them in the annotations folder and configure accuracyTracker.py
2. Put your images to be labeled in the /images folder
3. Run yolo.py, then accuracyTracker.py (if wanting to find accuracy)