测试时适应遇上图像增强:通过不确定性感知的 Logit 切换提升准确率
Test-time Adaptation Meets Image Enhancement: Improving Accuracy via Uncertainty-aware Logit Switching
AI总结:
针对分布偏移下测试时适应(TTA)中高不确定性预测降低准确率的问题,提出TECA方法,结合图像增强模型与分类器并引入logit切换机制输出低不确定性预测,有效提升了TTA方法的准确率。
摘要:
深度神经网络在各种计算机视觉应用中取得了显著的成功。然而,当数据分布在训练和测试之间发生偏移时,会出现准确率下降的问题。作为该问题的解决方案,测试时适应(Test-time Adaptation, TTA)因其实用性而得到了广泛研究。尽管 TTA 方法通过在测试时更新模型来提高分布偏移下的准确率,但已知使用高不确定性的预测会降低准确率。由于输入图像是分布偏移的根源,我们将增强输入图像的新视角引入 TTA 方法中,以降低预测的不确定性。我们假设,增强输入图像能够降低预测的不确定性并提高 TTA 方法的准确率。基于我们的假设,我们提出了一种新颖的方法:测试时增强器与分类器适应(Test-time Enhancer and Classifier Adaptation, TECA)。在 TECA 中,分类模型与图像增强模型相结合,后者将输入图像转换为对识别友好的图像,并且这些模型通过现有的 TTA 方法进行更新。此外,我们发现来自增强图像的预测并不总是具有比来自原始图像的预测更低的不确定性。因此,我们提出了 logit 切换,它比较这些预测的不确定性度量并输出较低不确定性的预测。在我们的实验中,我们用各种 TTA 方法评估了 TECA,并表明尽管没有超参数且参数开销极小,TECA 仍能降低预测的不确定性并提高 TTA 方法的准确率。
英文摘要:
Deep neural networks have achieved remarkable success in a variety of computer vision applications. However, there is a problem of degrading accuracy when the data distribution shifts between training and testing. As a solution of this problem, Test-time Adaptation~(TTA) has been well studied because of its practicality. Although TTA methods increase accuracy under distribution shift by updating the model at test time, using high-uncertainty predictions is known to degrade accuracy. Since the input image is the root of the distribution shift, we incorporate a new perspective on enhancing the input image into TTA methods to reduce the prediction's uncertainty. We hypothesize that enhancing the input image reduces prediction's uncertainty and increase the accuracy of TTA methods. On the basis of our hypothesis, we propose a novel method: Test-time Enhancer and Classifier Adaptation~(TECA). In TECA, the classification model is combined with the image enhancement model that transforms input images into recognition-friendly ones, and these models are updated by existing TTA methods. Furthermore, we found that the prediction from the enhanced image does not always have lower uncertainty than the prediction from the original image. Thus, we propose logit switching, which compares the uncertainty measure of these predictions and outputs the lower one. In our experiments, we evaluate TECA with various TTA methods and show that TECA reduces prediction's uncertainty and increases accuracy of TTA methods despite having no hyperparameters and little parameter overhead.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。