Fig. 1. Structure of the segmentation network based on deep supervision and U-Net

Fig. 2. Growth curve of the SC cells

Fig. 3. Segmentation results of different models. (a) Image 1; (b) image 2; (c) image 3

Fig. 4. FLIM images, distribution curves and statistical values of yeast cells at different ages. (a) FLIM images of t_m; (b) FLIM images of a₂; (c) distribution curve of t_m、a₂ and cross-sectional area; (d) statistical average value of t_m、a₂ and cross-sectional area

Fig. 5. Visualization results of t-SNE method. (a) t_m map; (b) a₂ map; (c) t_m and a₂ maps

Fig. 6. Clustering results and data distribution for difference feature input。(a) Input features are t_m and a₂; (b) two-dimensional feature distribution at 6 h; (c) two-dimensional feature distribution at 24 h; (d) two-dimensional feature distribution at 72 h; (e) input feature is t_m, a₂ and cross-sectional area; (f) three-dimensional feature distribution at 6 h; (g) three-dimensional feature distribution at 24 h; (h) three-dimensional feature distribution at 72 h

Table 1. Segmentation results of different segmentation models

Table 2. Proportion of the number of cells in different clusters in different input characteristics