Paper Title
Broiler Disease Detection using Yolov4

Abstract
Deep learning techniques have been highly potent in solving business problems related to Computer Vision and Natural Language Processing. It covers a broad range of applications and one of them is a disease detection tool for poultry health care management. Poultry farms are prone to various types of infection and if left alone, these infections can spread rapidly throughout the farm, resulting in the death of many birds and economic impact on the farm owners. A recent study showcased a tool that can identify two broiler diseases by analyzing its droppings. A deep convolutional neural network was used to achieve the goal. Our study expands on this initial work, with an additional disease to analyze and the use of a state-of-the-art object detection algorithm. We employed YOLOv4 for the task and from the experiment results, our model’s prediction time is much faster, at 23 milliseconds with a mean average precision of 87.5%. The experiment also involved training a scaled-down version of YOLOv4 for on-device computation. Keywords - Object Detection, Yolov4, Broiler Dropping Classification.