Fine-grained Metrics for Point Cloud Semantic Segmentation
This work addresses evaluation biases in point cloud segmentation for researchers, but it is incremental as it focuses on improving metrics rather than developing new segmentation methods.
The paper tackled the problem of biases in point cloud semantic segmentation evaluation due to category and size imbalances by proposing fine-grained mIoU and mAcc metrics, which were tested on three datasets to provide richer statistical information and reduce bias towards large objects.
Two forms of imbalances are commonly observed in point cloud semantic segmentation datasets: (1) category imbalances, where certain objects are more prevalent than others; and (2) size imbalances, where certain objects occupy more points than others. Because of this, the majority of categories and large objects are favored in the existing evaluation metrics. This paper suggests fine-grained mIoU and mAcc for a more thorough assessment of point cloud segmentation algorithms in order to address these issues. Richer statistical information is provided for models and datasets by these fine-grained metrics, which also lessen the bias of current semantic segmentation metrics towards large objects. The proposed metrics are used to train and assess various semantic segmentation algorithms on three distinct indoor and outdoor semantic segmentation datasets.