Enhancing Real-Time Object Detection Using YOLO-V7 Algorithm

The general objective of this study is to propose the methodology to improve the real-time object detection through the YOLO-V7 algorithm. The YOLO-V7 model is optimized in ways such as data augmentation enhancing the model, model pruning reducing complexities, hyperparameter tuning improving the generalizability, and post-processing optimization enhancing the model speed. The improved model has been proved to obtain the 88 % precision as shown in the experimental researches. 7 %, a recall of 86 % of materials taken for use and a scoring average of 64%. 50%, an AP of 85%, a median average precision (mAP) of 84%, and a mean average precision (mAP) of 87. By achieving real-time object detection with an efficiency of 6% and average inference time of 18ms, these upgrades are important for creating better object detection systems in a variety of contexts and uses. Keywords: Object detection, YOLO V7, real-time implementation, data augmentation, model quantization, hyperparameter optimization, post-processing optimization.