Application of extreme learning machine with 3D image description for medical objects segmentation
Medical image segmentation has many applications in health care industry. This project aims at applying the newly developed learning algorithm - Extreme Learning Machine (ELM) to 3D-medical images to segment liver image based on texture features. Further to liver segmentation, a comparison is made b...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/49841 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Medical image segmentation has many applications in health care industry. This project aims at applying the newly developed learning algorithm - Extreme Learning Machine (ELM) to 3D-medical images to segment liver image based on texture features. Further to liver segmentation, a comparison is made between direct ELM segmentation and ELM segmentation with adaboosting on liver tumor segmentation. A study of Active Shape Model (ASM) is carried out to regulate the shape obtained in liver segmentation.
Multiple texture features for the medical image are extracted from 3D CT images before being trained and tested using ELM. A variety of methods are applied to measure the performance. Several pre-processing and post-processing methods such as alignment, morphological operations are used to improve the accuracy in classification. Active Shape Model allows the shape of testing data which is represented by a series of landmarks to change within constrain of the training model. The contour generated from ELM segmentation is used to obtain the initial landmarks for ASM. |
|---|