Laparoscopic Image-Guided System Based on Multispectral Imaging for the Ureter Detection
| Autor: | Zhu Jun, Enmin Song, Chih-Cheng Hung, Hong Liu, Yu Feng |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
Laparoscopic surgery
General Computer Science Endoscope image-guided Computer science Urinary system medicine.medical_treatment Multispectral image Connective tissue 02 engineering and technology 01 natural sciences Image (mathematics) Ureter Peritoneum 0202 electrical engineering electronic engineering information engineering medicine multispectral imaging General Materials Science Computer vision Segmentation business.industry 010401 analytical chemistry General Engineering 020206 networking & telecommunications Frame rate ureter injury 0104 chemical sciences medicine.anatomical_structure Artificial intelligence Ureter injury Endoscope system lcsh:Electrical engineering. Electronics. Nuclear engineering business lcsh:TK1-9971 ureter detection |
| Zdroj: | IEEE Access, Vol 7, Pp 3800-3809 (2019) |
| ISSN: | 2169-3536 |
| Popis: | The iatrogenic ureter injury is a common medical negligence in the gynecology, abdominal, and urinary surgeries. Anatomically, the ureter is covered by peritoneum and connective tissue, and the doctor cannot observe it directly in surgery. The ureter injury may cause significant complications for patients and medical disputes. It is important to indicate the ureter position for aided surgery of the doctor. To provide ureter position for doctors in the laparoscopic surgery, we design an image-guided endoscope system that includes a novel endoscopic video system with a visible-light camera and an infrared camera. The visible-light camera is to capture the coeliac image and the infrared camera is to capture the ureter position, simultaneously. To extract accurate ureter position in the infrared image, we also propose a self-adaptive threshold segmentation algorithm to extract the real ureter position as accurately as possible. The self-adaptive threshold and scattering factor are taken in to full account for the ureter segmentation. In addition, the scattering property of light is also discussed to choose the optimal light. Finally, we design and develop the image-guided endoscope system, and experiment it on the animal. The experimental results demonstrate that the proposed image-guided endoscope system achieves 93.8% and 90.6% in terms of true positive rate and positive predictive value, respectively. The processing speed of the proposed algorithm can reach about 165 frames per second (f/s), and the frame rate is far faster than the frame rate (30 f/s) of the traditional endoscope system. The accuracy and processing ability of the system can satisfy the clinical demand. The iatrogenic ureter injury may be decreased when the surgeons perform the operations with the ureter position displayed in real time. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|