预测图像分类器在失真下的可靠性，提升模型质量控制能力。
Predicting the Reliability of an Image Classifier under Image Distortion
- 构建失真等级与可靠/不可靠标签的训练集，用机器学习预测可靠性。
- 在六个主流数据集上，性能显著优于多个基线方法。
- 采用高斯过程重平衡数据集，解决训练样本严重不均衡问题。
在图像分类任务中，深度学习模型对图像失真敏感，输入图像失真时准确率会显著下降。若分类器在失真图像上的准确率高于用户设定阈值，则认为其为“可靠”。为实现质量控制，需预测分类器在特定失真水平下的可靠性。本文提出一种方法：构建包含失真等级及其“可靠”或“不可靠”标签的训练集，训练一个称为“失真分类器”的机器学习模型，以分类未见的失真等级。由于训练集高度不平衡，学习有效的失真分类器极具挑战。为此，我们提出基于高斯过程的方法来重平衡训练集。大量实验表明，该方法在六个流行图像数据集上显著优于多个基线模型。
原文摘要 · Abstract (English)
In image classification tasks, deep learning models are vulnerable to image distortions i.e. their accuracy significantly drops if the input images are distorted. An image-classifier is considered "reliable" if its accuracy on distorted images is above a user-specified threshold. For a quality control purpose, it is important to predict if the image-classifier is unreliable/reliable under a distortion level. In other words, we want to predict whether a distortion level makes the image-classifier "non-reliable" or "reliable". Our solution is to construct a training set consisting of distortion levels along with their "non-reliable" or "reliable" labels, and train a machine learning predictive model (called distortion-classifier) to classify unseen distortion levels. However, learning an effective distortion-classifier is a challenging problem as the training set is highly imbalanced. To address this problem, we propose a Gaussian process based method to rebalance the training set. We conduct extensive experiments to show that our method significantly outperforms several baselines on six popular image datasets.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。