微創手術器械偵測

Abstract

隨著科技的進步，醫療科技也蓬勃發展，近年來，微創腹腔鏡手術的興盛，使醫生在手術過程中，只需要微小的手術傷口就能進行治療，但是醫生在手術過程全都是看著螢幕操作，遠近沒有立體感，需要非常純熟的技術才能順利操作。為了能讓醫生更明確知道手術器械位置，並達到提醒手術器械與手術部位相對位置的功用，避免不必要的傷害，因此透過專題研究「器械偵測」，來輔助醫生的手術進行。 微創手術進行時，經由小傷口將鏡頭及器械分別深入手術部位，可拍攝到器械及欲手術的部位，系統做偵測並即時框出器械，提供醫生更有效率的判斷器械位置，有效避免因疲勞誤判而造成的傷害。本研究主要著重在器械的形狀、顏色、狀態，運用Adaboost之Haar-like、HSV色彩空間轉換、Motion Detection等方法做影像的物件偵測，來輔助醫生的手術進行。 目前已有許多偵測技術應用在行車上，若此器械偵測技術能廣泛應用在手術中，必定能使手術過程更加順利，也是醫生及患者的一大福音。

With the advancement of technology, medical technology is booming in recent years, the rise of minimally invasive laparoscopic surgery. During the surgery, the doctor only needs a tiny incision to treat, in the surgical procedures. The doctor only needs to look at screens, but the far have no dimensional sense. It needs very skillful technology to operate. In order to make doctors more aware of surgical instruments and the reminding function of the relative position of the surgical instruments and surgical site, to avoid unnecessary damage. Therefore the thematic study of “invasive surgical instruments", to aid in the doctor's surgery. During the minimally invasive surgery, we through small wounds lenses and the equipment further into the surgical site, it can screen the parts of the operation .The system can detect and catch the instrument timely. To provide doctors determine the devices efficiently .It can effectively avoid the fatigue damage caused by miscalculation. This study focuses on the shape, color, and the condition of the device, using the Adaboost Haar-like, HSV color space conversion, and the Motion Detection method, such as object detection on visual image techniques to aid the doctor's surgery. There are many detection technologies that are set up on the vehicles, if the device detection technology can be widely used in surgery. It will be able to make the process more smoothly and be a great boon to doctors and patients.