The quality of training datasets for deep neural networks is a key factor contributing to the accuracy of resulting models. This effect is amplified in difficult tasks such as object detection. Dealing with errors in datasets is often limited to accepting that some fraction of examples is incorrect, estimating their confidence and assigning appropriate weights or ignoring uncertain ones during training. In this work, we propose a different approach. We introduce the Confident Learning for Object Detection (CLOD) algorithm for assessing the quality of each label in object detection datasets, identifying missing, spurious, mislabeled and mislocated bounding boxes and suggesting corrections. By focusing on finding incorrect examples in the training datasets, we can eliminate them at the root. Suspicious bounding boxes can be reviewed in order to improve the quality of the dataset, leading to better models without further complicating their already complex architectures. The proposed method is able to point out 99% of artificially disturbed bounding boxes with a false positive rate below 0.3. We see this method as a promising path to correcting popular object detection datasets.

翻译：深度神经网络训练数据集的质量是影响最终模型精度的关键因素之一，这一影响在目标检测等复杂任务中尤为显著。当前处理数据集误差的方法通常局限于接受部分示例存在错误的事实，通过估计其置信度并在训练过程中为其分配适当权重或忽略不确定样本。本研究提出了一种不同思路：我们引入面向目标检测的置信学习（CLOD）算法，用于评估目标检测数据集中每个标签的质量，识别缺失、虚假、错误标注及定位偏差的边界框，并提出修正建议。通过聚焦训练数据集中错误示例的定位，我们能够从根源上消除这些错误。可疑的边界框可被复核以提升数据集质量，从而在无需进一步复杂化本已精密的网络架构的前提下获得更优模型。该方法能以低于0.3%的假阳性率检出99%的人工扰动边界框。我们认为该方法为修正主流目标检测数据集提供了可行路径。