Describe what the paper have done, training srategy (supervised/unsupervised), system inputs and outputs, and application scenarios.
Very important.
Clarify the motivation behind the paper, and the comparison with prior arts.
- Using screenshots if neccessary. Directly ctrlC+V.
- Provide the network details.
Implementation details of contributions. Description of novel ideas.
If the paper introduces new datasets or metrics, please update the datasets_metrics.md.
provide the experimental results like:
nuScenes validation set:
- IoU: 29.83
- MIoU: 19.10
xx xx set:
- xx: xx
- xx: xx
Very important but not mandatory.