Single-stage object detection algorithms are a type of computer vision technique used to identify and locate objects within an image or video in real-time. Unlike their two-stage counterparts, which involve region proposal and object classification in separate steps, single-stage detectors directly predict bounding boxes and class probabilities in a single pass. One of the most well-known single-stage detection algorithms is YOLO (You Only Look Once). YOLO divides the input image into a grid and predicts bounding boxes, confidence scores, and class probabilities for each grid cell. This allows YOLO to efficiently process the entire image at once, making it fast and suitable for real-time applications.

While Yolo has had phenomenal success with its use case being extended to a wide variety of applications, it still needs to improve upon its accuracy as well as be versatile in all the different use cases. Hence we propose a method to integrate knowledge graphs and semantic analysis into a Yolo-like single stage detector algorithm to develop single stage detection to be easily used in different industries such as sports video analysis, anomaly detection, vqa-based reasoning problems, etc., The novelty of this project lies in the fact that while graph methods have tried to be integrated with single stage detection algorithms in the past, most of them had tried to develop convolutions with graph convolutional neural networks in order to handle graph structured data.None of these papers have tried to improve upon reasoning based attributes of the algorithm. Hence what we propose involves building a neuro-symbolic architecture where the deep learning part is used to extract important features (sensors of the machine) while the knowledge graph and semantic analysis part is used in reasoning based problem solving and also to improve adaptability to different use cases.