Fig 1.
Blender scene showcasing strawberry plant (top) alongside a vertical grow wall and a camera mounted on a track (bottom).
Fig 2.
Examples of synthetic images featuring both ripe and unripe strawberries, along with their corresponding masks.
Fig 3.
Examples of real test images featuring both ripe (top) and unripe (bottom) strawberries.
Fig 4.
Example of grayscaled real-world image (left) alongside corresponding segmentation mask (right).
Fig 5.
SwinUNet architecture, composed of encoder, decoder, bottleneck, and skip connections.
Table 1.
Model parameters and specifications.
Table 2.
DSC for 10,000 images on validation and test datasets.
Table 3.
DSC improvement as number of training images increases.