Tracking Pointer At Endoscopic Images For Telepointer Remote Guided
Nowadays, the telepointer landmark (TL) has become a crucial tool for guiding surgeons during surgery. To keep detect and track of the telepointer landmark (TL) at tissues, the current approach such as Template Matching, is applied. Unfortunately, due to changes in tissue characteristics such as con...
Saved in:
| Main Author: | |
|---|---|
| Format: | Undergraduates Project Papers |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/39865/1/EA18008_Muhamad%20Hamizi%20Zaidi%20Bin%20Mohd%20Jonhanis_Thesis%20-%20Muhamad%20Hamizi%20Zaidi.pdf http://umpir.ump.edu.my/id/eprint/39865/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Nowadays, the telepointer landmark (TL) has become a crucial tool for guiding surgeons during surgery. To keep detect and track of the telepointer landmark (TL) at tissues, the current approach such as Template Matching, is applied. Unfortunately, due to changes in tissue characteristics such as continuous movement, non-rigid, homogeneous texture, and varied illumination, the precision of the matched point is always bound in relying on a single hard decision. Several options are required to maintain track of actual activity based on the aspect of the tissues. The first problem is the detection of the cursor. Experts and surgeons are likely to overlook the cursor's location during surgery. In the traditional method, the pointer is difficult to detect with the naked eye. This is because the cursor’s colour is usually white and white has bad contrast to blood and tissue colour so it will confuse the expert in detecting the pointer or cursor. The second problem is that an expert must point the cursor or pointer to the mark location and manually follow the movement. The manipulation of the camera and the coordination between surgeon and assistant can be complex and require excellent skills from the doctor and the surgeon on site. This study aims to detect and track the cursor a system using computer vision technology. Therefore, the first aim is to detect the green cursor or pointer since the colour of cursor used is green colour. The second aim is to track the mark location that the pointer or cursor has marked. In this project, for cursor detection, the proposed colour tracking for cursor detection is RGB Colour Space, HSV, Colour Space and K-Means Clustering. For cursor tracking, the proposed cursor tracking is Blob Detection (Colour Feature) and Feature Point Tracking. The experimental results for the proposed system cursor detection system show that HSV is highlighted with more positive aspects than other methods. Therefore, more suitable for green colour cursor detection and will be used in Blob Detection (Colour Feature) method. For the cursor tracking methods, the experimental results show that the system using Blob Detection (Colour Feature) and Features Point Matching is a success where the systems are able to detect and track the cursor or mark location. Both methods have 100% accuracy in detecting and tracking the cursor. The proposed future cursor detection and tracking system should be capable of applying the proposed algorithm in a more complex environment and various input data. |
|---|