Text this: Comparative classification performance using VGG, ResNet, and ViT models, with each model applied in two different training approaches on segmented cells: trained from scratch and fine-tuned.