The SEE-AI Project Dataset

About Dataset

Introducing the SEE-AI Project Dataset, a game-changer in the realm of small bowel capsule endoscopy (CE) research. Our dataset features over 18,000 meticulously annotated images extracted from CE videos, aiming to advance automated reading techniques for small bowel CE examinations.

About the File

The Robotics deal file is a comprehensive collection of 18,481 images extracted from 523 small bowel capsule endoscopy videos. These images have been meticulously annotated, with 23,033 disease lesions and 6,161 normal mucosa images, all presented in YOLO format. Our aim is to provide a robust and diverse dataset that serves as a foundation for the development of automated or assisted reading techniques for small bowel CE.

Challenges in Small Bowel CE Examination

Challenges abound in accurately identifying lesions from CE images, akin to expert physicians. Our research, detailed in the paper “Small Bowel Capsule Endoscopy Examination with Object Detection Artificial Intelligence Model: The SEE-AI Project,” explores these challenges. Leveraging YOLO v5, we’ve crafted annotations and developed an object detection model to tackle these obstacles head-on.

Future Prospects

Looking ahead, we believe in exploring innovative annotation methods to refine AI models’ accuracy and efficiency in small bowel CE examination. The SEE-AI Project Dataset serves as a vital resource for researchers, developers, and medical professionals striving to revolutionize small bowel CE object detection.

Conclusion

Join us in leveraging this dataset to drive further research and innovation, ultimately enhancing the precision and speed of small bowel CE exams. Together, we can usher in a new era of AI-powered healthcare, benefiting patients worldwide.