Current state-of-the-art two-stage models on instance segmentation task suffer from several types of imbalances. In this paper, we address the Intersection over the Union (IoU) distribution imbalance of positive input Regions of Interest (RoIs) during the training of the second stage. Our Self-Balanced R-CNN (SBR-CNN), an evolved version of the Hybrid Task Cascade (HTC) model, brings brand new loop mechanisms of bounding box and mask refinements. With an improved Generic RoI Extraction (GRoIE), we also address the feature-level imbalance at the Feature Pyramid Network (FPN) level, originated by a non-uniform integration between low- and high-level features from the backbone layers. In addition, the redesign of the architecture heads toward a fully convolutional approach with FCC further reduces the number of parameters and obtains more clues to the connection between the task to solve and the layers used. Moreover, our SBR-CNN model shows the same or even better improvements if adopted in conjunction with other state-of-the-art models. In fact, with a lightweight ResNet-50 as backbone, evaluated on COCO minival 2017 dataset, our model reaches 45.3% and 41.5% AP for object detection and instance segmentation, with 12 epochs and without extra tricks. The code is available at https://github.com/IMPLabUniPr/mmdetection/tree/sbr_cnn

翻译：当前最先进的两阶段实例分割模型在训练过程中存在多种类型的不平衡问题。本文主要解决第二阶段训练中正样本感兴趣区域（RoI）的交并比（IoU）分布不平衡问题。我们提出的自平衡R-CNN（SBR-CNN）作为混合任务级联（HTC）模型的进化版本，引入了全新的边界框与掩膜精化循环机制。通过改进的通用RoI特征提取模块（GRoIE），我们同时解决了特征金字塔网络（FPN）层级上由于骨干网络低层与高层特征非均匀融合导致的特征级不平衡问题。此外，架构头部采用全卷积FCC方案重新设计，在减少参数量的同时，增强了目标任务与网络层之间的关联性。值得关注的是，SBR-CNN模型与其他先进模型联用时仍能保持同等甚至更优的提升效果。实验表明，在采用轻量级ResNet-50作为骨干网络、COCO minival 2017数据集上仅训练12轮且未使用任何额外技巧的情况下，我们的模型在目标检测与实例分割任务上分别达到45.3%和41.5%的平均精度（AP）。代码已开源：https://github.com/IMPLabUniPr/mmdetection/tree/sbr_cnn