Rapid identification of victims is vital for efficient rescue operations in the aftermath of natural disasters like earthquakes. This study evaluates the performance of two advanced object detection models, YOLOv7 and YOLOv8, for victim detection in disaster scenarios. Both models were trained on a dataset simulating post-disaster environments to recognize human bodies among debris and challenging conditions. The Simulated Disaster Victim (SDV1 and SDV2) dataset is an openly available dataset aimed at training and testing DL algorithms in the seek for identifying disaster victims. The experimental results showed an accuracy of 58% for YOLOv7 and a significantly improved accuracy of 81% for YOLOv8, indicating a notable advancement in the latter's ability to detect victims accurately. This comparative analysis not only highlights the superior performance of YOLOv8 but also explores the strengths and limitations of YOLOv7 in terms of detection accuracy, precision, and recall. The findings underscore the potential of employing YOLOv8 for real-time disaster response systems, where quick and reliable victim identification can save lives. Future work could focus on enhancing the model's robustness in diverse scenarios and integrating additional sensor data for improved detection with the latest versions of the YOLO series.