Skip to main content
. Author manuscript; available in PMC: 2025 Jun 9.
Published in final edited form as: Comput Biol Med. 2024 Jul 7;179:108846. doi: 10.1016/j.compbiomed.2024.108846

Table 2.

F-measure values for Cellpose-CPPA and CellProfiler cytoplasm segmentation of NAD(P)H images of quiescent T cells, activated T cells, and MCF7 cells. F-measure values are computed for each identified cell via POSEA. Values are reported as mean (± standard deviation), n = 982 quiescent T cells, 228 activated T cells, and 335 MCF7 cells. From the subsequent statistical analysis, it was found that the F-measure for the CellProfiler method is significantly lower (p < 2e-16) compared to the CPPA threshold methods. In addition, within the CPPA thresholding methods, no significant differences were observed except for the mean thresholding method applied to quiescent T cells, which showed a significantly lower F-measure (p < 2e-16) compared to other thresholding methods. This aligns with the analysis presented in Table 1.

Segmentation Algorithm F-measure

Quiescent T cells Activated T cells MCF7 cells

CellProfiler 0.59 (±0.23) 0.55 (±0.21) 0.46 (±0.29)
CPPA-isodata 0.83 (±0.09) 0.76 (±0.13) 0.72 (±0.21)
CPPA-li 0.83 (±0.09) 0.75 (±0.14) 0.72 (±0.21)
CPPA-mean 0.80 (±0.09) 0.75 (±0.14) 0.72 (±0.21)
CPPA-otsu 0.83 (±0.09) 0.76 (±0.14) 0.72 (±0.21)
CPPA-triangle 0.84 (±0.08) 0.76 (±0.14) 0.69 (±0.20)
CPPA-yen 0.83 (±0.09) 0.75 (±0.13) 0.72 (±0.21)