A simple comparison and example implementation of object detection algorithms covering VGG and YOLO v8 for comparison of their functionality and as a starting point for a detection project. The implementation expects images to be labelled using Label Studio and their XML file exports.
Due to the earlier implementation of the VGG algorithm the labelled information has to be transformed into a CSV file format. In order to transpose the data into the expected format run the preprocess.py function. It will create an index of the given dataset.
The code is based in parts on the yolov8 example from the keras-io repository and adapted for these purposes. That code in specific is distributed under the Apache License 2.0. You may obtain a copy of the License at apache.org or in the LICENSE file in the root directory of its repository.