Before we proceed further, we'll understand the task of object detection. The standard definition from Wikipedia goes like, The task of object detection consists of two things - th… See more

The architecture consists of various parts, broadly they are - The input which comes first and it is basically what we've as our set of training images which will be fed to the network - t… See more

The authors initially considered CSPResNext50, CSPDarknet53 and EfficientNet-B3 as the backbone networks. Finally, after a lot of testing and experiment… See more

The neck is the part where feature aggregation takes place. It collects feature maps from the different stages of the backbone then mixes and combines them to prepare t… See more

The main function here is locating bounding boxes and performing classification. The bounding box co-ordinates (x, y, height and width) as well as scores are detected. Here the x & y c… See more