MitoVis is a novel visual analysis system specifically designed for end-to-end neuronal mitochondria morphology analysis on light (confocal) microscopy data. MitoVis allows fast and accurate neuronal structure segmentation and mitochondria detection via interactive fine-tuning of the deep learning model. To achieve this, MitoVis provides a novel visualization-guided data suggestion scheme combined with a pretrained encoder using contrastive loss and tracing-based neurite-level classification for fast and accurate fine-tuning of the model.