The objective of the Object Detection track of the Open Images Challenge
2019 hosted by Kaggle was to detect multiple objects within an image,
and identify the class they belonged to. To achieve this, we have used
two different models, YOLOv3 and RetinaNet, implemented on Keras,
Both models have their own strengths, speed and accuracy, respectively.
The models have been trained on a subset of the original dataset due to
time and computational constraints.