Geometric Feature Extraction for Identification and Classification of the Overlapping Cells for Leukaemia
Overlapping cell identification and classification of microscopic blood cell image is proposed to increase the accuracy of the number of automated counting and the percentage of the overall accuracy. The accurate identification of overlapping cells can increase the accuracy of cell counting for dia...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/26606/1/Kiu%20Siew%20ft.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Overlapping cell identification and classification of microscopic blood cell image is proposed to increase the accuracy of the number of automated counting and the percentage of the overall accuracy. The accurate identification of overlapping cells can increase the
accuracy of cell counting for diagnosing Leukaemia disease. In the proposed method, the percentage of average accuracy for identifying overlapping cells had reached 98%. The overlapping cells had classified into different classes based on overlapping degree and the
number of overlapping cells. The proposed method can successful segmentation white blood cell from the overlapping cells with 70%. The average percentage of cell counting had tested by the proposed method. Extended-Minima Transform Watershed Segmentation
Algorithm had successful increasing 36.89% of accuracy to WBC segmentation. It had reduced 46.12% of over-segmentation problem. Tests of cell counting in between the two stages, which are before and after for the process of identification and classification of
overlapping cells. The average percentage of total cell count is 83.31%, the average percentage of WBC counting is 84.8% and the average percentage of RBC counting is
60.55%. |
|---|