This paper presents an innovative approach for detecting bleeding and non-bleeding instances within the gastrointestinal (GI) tract using Wireless Capsule Endoscopy (WCE) images. Leveraging the capabilities of YOLOv5, a state-of-the-art object detection algorithm, our methodology involves training the model on a carefully annotated dataset of WCE images categorized as bleeding or non-bleeding. The trained model is then employed for real-time detection of bleeding and non-bleeding regions in new WCE images. Key contributions include the adoption of YOLOv5 for robust object detection, meticulous dataset annotation, real-time detection capabilities, comprehensive performance evaluation metrics, and a discussion on the implications for clinical practice. The integration of our approach holds promise for enhancing early diagnosis and intervention in cases of gastrointestinal bleeding, showcasing the adaptability and efficacy of YOLOv5 in medical imaging contexts.