Interactive medical image segmentation with seed precomputation system: Data from the Osteoarthritis Initiative
Although interactive segmentation helps lower the degree of human intervention, further upgrade to increase the efficiency and intuitiveness of interactive segmentation method remains requisite. Invariably, newly acquired medical images are not segmented instantly by radiologists due to heavy work l...
保存先:
| 主要な著者: | , , , , |
|---|---|
| フォーマット: | Conference or Workshop Item |
| 出版事項: |
2015
|
| 主題: | |
| オンライン･アクセス: | http://eprints.utm.my/id/eprint/59331/ http://dx.doi.org/10.1109/IECBES.2014.7047510 |
| タグ: |
タグ追加
タグなし, このレコードへの初めてのタグを付けませんか!
|
| 要約: | Although interactive segmentation helps lower the degree of human intervention, further upgrade to increase the efficiency and intuitiveness of interactive segmentation method remains requisite. Invariably, newly acquired medical images are not segmented instantly by radiologists due to heavy work load. Therefore, improvement can be made by capitalizing on the time interval between image acquisition and image segmentation. We proposed to pre-generate non-cartilage seeds during this time interval so the labeling process can be simplified. This enhanced interactive segmentation decreases processing time required by a radiologist without compromising the quality of original segmentation quality. Future works should focus on developing automatic pre-generation of different types of cartilage labels. |
|---|