Complex YOLOv3 VS DeepLabV3_Plus-Tensorflow2.0 Deep Learning Experiences #16
This video illustrates how much the semantic segmentation is dependent on the database. CNN algorithms like YOLOv3 are also, but the databases (especially for pedestrians) are much more complete and this type of algorithm is less sensitive to variations in brightness (despite some false positives) in comparison with the segmentation algorithm sematic especially thanks to the many filters used