TrueSight

A YOLOv3 based object detection + voice feedback model.

Steps to run:

1. The code was run on Windows 10 and the python version used is 3.8.5
2. Download dependencies; OpenCV, os, numpy and gtts (using pip)
3. An internet connection is required for gtts to function
4. Download the yolov3 weights from the link: https://pjreddie.com/media/files/yolov3.weights
5. yolov3 is a pre trained model, these weights are a result of training on the MS-COCO dataset: https://cocodataset.org/#download
6. Place the weights folder in the same directory as the Python file
7. Run the following command: python trueSight_yolo.py

• The processed image will then be displayed in its own window, it will also be saved in the directory as object-detection.jpg
• Once this window is closed, the generated mp3 file will begin playing.
• One can analyze their own images by saving it in the images folder and altering the file path in line 21 of the python file.