Improving the Intra-class Long-tail in 3D Detection via Rare Example Mining

  title={Improving the Intra-class Long-tail in 3D Detection via Rare Example Mining},
  author={Chiyu Max Jiang and Mahyar Najibi and C. Qi and Yin Zhou and Drago Anguelov},
. Continued improvements in deep learning architectures have steadily advanced the overall performance of 3D object detectors to levels on par with humans for certain tasks and datasets, where the overall performance is mostly driven by common examples. However, even the best performing models suffer from the most naive mistakes when it comes to rare examples that do not appear frequently in the training data, such as vehicles with irregular geometries. Most studies in the long-tail literature… 

Figures from this paper

AdaptiveShape: Solving Shape Variability for 3D Object Detection with Geometry Aware Anchor Distributions

AdaptiveShape, a new fast LiDAR-camera fusion method based on 2D bounding box camera detections which is available in many processing pipelines, is introduced and extended to account for temporal data and improve learning of complex object movements.



Overcoming Classifier Imbalance for Long-Tail Object Detection With Balanced Group Softmax

  • Yu LiTao Wang Jiashi Feng
  • Computer Science
    2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2020
This work provides the first systematic analysis on the underperformance of state-of-the-art models in front of long-tail distribution and proposes a novel balanced group softmax (BAGS) module for balancing the classifiers within the detection frameworks through group-wise training.

Equalization Loss for Long-Tailed Object Recognition

This work proposes a simple but effective loss, named equalization loss, to tackle the problem of long-tailed rare categories by simply ignoring those gradients for rare categories, and wins the 1st place in the LVIS Challenge 2019.

Feature Space Augmentation for Long-Tailed Data

This work presents a novel approach to address the long-tailed problem by augmenting the under-represented classes in the feature space with the features learned from the classes with ample samples, and decomposes the features of each class into a class-generic component and aclass-specific component using class activation maps.

DropLoss for Long-Tail Instance Segmentation

It is demonstrated that the rare categories are heavily suppressed by correct background predictions, which reduces the probability for all foreground categories with equal weight, and develops DropLoss – a novel adaptive loss to compensate for this imbalance without a trade-off between rare and frequent categories.

Not All Labels Are Equal: Rationalizing The Labeling Costs for Training Object Detection

This work proposes a unified frame-work for active learning, that considers both the uncertainty and the robustness of the detector, ensuring that the network performs well in all classes, and leverages auto-labeling to suppress a potential distribution drift while boosting the performance of the model.

Classification Calibration for Long-tail Instance Segmentation

This report investigates the performance drop phenomenon of state-of-the-art two-stage instance segmentation models when processing extreme long-tail training data based on the LVIS dataset, and finds a major cause is the inaccurate classification of object proposals.

Seesaw Loss for Long-Tailed Instance Segmentation

This work proposes Seesaw Loss to dynamically re-balance gradients of positive and negative samples for each category, with two complementary factors, i.e., mitigation factor and compensation factor, which obtains significant gains over Cross-Entropy Loss and achieves state-of-the-art performance on LVIS dataset without bells and whistles.

Training Region-Based Object Detectors with Online Hard Example Mining

OHEM is a simple and intuitive algorithm that eliminates several heuristics and hyperparameters in common use that leads to consistent and significant boosts in detection performance on benchmarks like PASCAL VOC 2007 and 2012.

Decoupling Representation and Classifier for Long-Tailed Recognition

It is shown that it is possible to outperform carefully designed losses, sampling strategies, even complex modules with memory, by using a straightforward approach that decouples representation and classification.

Self Supervision to Distillation for Long-Tailed Visual Recognition

It is shown that soft label can serve as a powerful solution to incorporate label correlation into a multi-stage training scheme for long-tailed recognition, as well as a new distillation label generation module guided by self-supervision.