DyConfidMatch: Dynamic Thresholding and Re-sampling for 3D Semi-supervised Learning
This addresses data imbalance issues in 3D semi-supervised learning, which is important for applications like autonomous driving and robotics, though it appears to be an incremental improvement over existing SSL methods.
The paper tackles data imbalance in 3D semi-supervised learning by proposing a method that uses dynamic thresholding based on class-level confidence and re-sampling to improve representation of underrepresented classes. It achieves state-of-the-art results in 3D classification and detection tasks.
Semi-supervised learning (SSL) leverages limited labeled and abundant unlabeled data but often faces challenges with data imbalance, especially in 3D contexts. This study investigates class-level confidence as an indicator of learning status in 3D SSL, proposing a novel method that utilizes dynamic thresholding to better use unlabeled data, particularly from underrepresented classes. A re-sampling strategy is also introduced to mitigate bias towards well-represented classes, ensuring equitable class representation. Through extensive experiments in 3D SSL, our method surpasses state-of-the-art counterparts in classification and detection tasks, highlighting its effectiveness in tackling data imbalance. This approach presents a significant advancement in SSL for 3D datasets, providing a robust solution for data imbalance issues.