| 研究課題/領域番号 |
16F16108
|
|---|
| 研究機関 | 九州大学 |
|---|
研究代表者 |
橋爪 誠 九州大学, 医学研究院, 教授 (90198664)
|
|---|
| 研究分担者 |
JANG JONGSEONG 九州大学, 医学研究院, 外国人特別研究員
|
|---|
| 研究期間 (年度) |
2016-07-27 – 2019-03-31
|
| キーワード | Surgical navigation / Medical image anaysis / Object tracking / Image fusion |
|---|
| 研究実績の概要 |
In this year, I achived in implementation of the components of the navigation for robot assisted (da Vinci assisted) radical prostectomy. I analyzed medical software that has been utilized by surgeons of Kyushu University Hospital, and inserted the navigation functional component (Optical tracker, Modified registration module, Ultrasound deviceand etc.).
In addition, special function to show the boundaris between bladder, prostate, and to guide the surgical instruments into that was implemented to help the surgeon. This function can show the the correct path to cut the boundary between the prostate and bladder so that prevent mis-recognition about the their anatomical position and shape.
Moreover, forceps tracking component was fundamentally implemented. Convetionally, forceps tracking techniques cannot show their performance in real surgery because their achivement was accoplished in their lab. I had to solve this problem in real surgical situation, I achived a sort of effective method for it using deep neural network.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
There was no big problem in installation of the optical tracking system, modification of software, addition of functional modules. However, there were some difficulties in da Vinci forceps tracking. There was no way to analyze forceps kinematics because there was no information about the da Vinci mechanical design. I had to extract only in endoscopic movie. Conventional forceps tracking technologies could not help in the real surgery. However, I solved this problem a little, and I expect it can be available when using my novel method.
|
| 今後の研究の推進方策 |
Forceps tracking is still slow to applied to navigation system. Through optimization of algorithm, I would like to imporve tacking speed. There are some proposed deep neural network like algorithm to detecting object in a movie. I'll try them and apply the optimized method. In addition, through the two endoscopic display(left-right), three dimensional position of the forceps will be calculated by using triangulation technique. This procedure can be achived by camera calibration of da Vinci. There is one residual da Vinci endoscope in this lab so that it is possible perform experiment so long time. To guarante accuracy of three position of the forceps, I have to many chance to measure its accuracy through many chance of experiment.
|
