Key references: (Yuan et al., 2019; Ibrahim et al., 2018; Zhao et al., 2019; Cordts et al., 2016; Li et al., 2018)
UNet from scratch
Build UNet in PyTorch from DoubleConv, Down, Up and OutConv blocks, train it on FoodSeg103 with Dice and cross-entropy loss, and evaluate it with IoU.
References
- Cordts, M., Omran, M., Ramos, S., Rehfeld, T., Enzweiler, M., et al. (2016). The Cityscapes Dataset for Semantic Urban Scene Understanding. arXiv [cs.CV].
- Ibrahim, M., Vahdat, A., Ranjbar, M., Macready, W. (2018). Semi-Supervised Semantic Image Segmentation with Self-correcting Networks. arXiv [cs.CV].
- Li, Y., Shi, J., Lin, D. (2018). Low-Latency Video Semantic Segmentation. arXiv [cs.CV].
- Yuan, Y., Chen, X., Chen, X., Wang, J. (2019). Segmentation Transformer: Object-Contextual Representations for Semantic Segmentation. arXiv [cs.CV].
- Zhao, S., Wang, Y., Yang, Z., Cai, D. (2019). Region mutual information loss for semantic segmentation. arXiv [cs.CV].