PulseAugur
LIVE 16:00:51
research · [2 sources] ·
0
research

New loss functions improve multi-class semantic segmentation accuracy

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 2 sources

Researchers have developed new loss functions for multi-class semantic segmentation that address both instance and class imbalance issues. By extending instance-sensitive losses like blob loss and CC loss to multi-class settings, these methods improve the training signal for rare classes. The proposed techniques, including a one-vs-rest decomposition and integrated inverse-size weighting, demonstrated performance gains on the BraTS-METS 2025 dataset, enhancing metrics such as foreground Dice and Panoptic Quality. AI

Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →

IMPACT Improves segmentation accuracy for rare classes, potentially benefiting medical imaging and other applications with imbalanced datasets.

RANK_REASON This is a research paper detailing new methods for semantic segmentation loss functions.

Read on arXiv cs.CV →

COVERAGE [2]

  1. arXiv cs.CV TIER_1 · Soumya Snigdha Kundu, Florian Kofler, Marina Ivory, Hendrik Moller, Jonathan Shapey, Tom Vercauteren ·

    Instance Awareness of Multi-class Semantic Segmentation Loss Functions

    arXiv:2604.24276v1 Announce Type: new Abstract: Instance-sensitive losses for semantic segmentation such as blob loss and CC loss were designed to address instance imbalance, ensuring small lesions generate the same gradient as large ones, but operate only on single-class segment…

  2. arXiv cs.CV TIER_1 · Tom Vercauteren ·

    Instance Awareness of Multi-class Semantic Segmentation Loss Functions

    Instance-sensitive losses for semantic segmentation such as blob loss and CC loss were designed to address instance imbalance, ensuring small lesions generate the same gradient as large ones, but operate only on single-class segmentation. In multi-class settings, class imbalance …

RELATED ENTITIES

RELATED TOPICS