我无法相信TTA并不更好：当测试时间增强伤害医学图像分类
I Can't Believe TTA Is Not Better: When Test-Time Augmentation Hurts Medical Image Classification
AI总结：
本文通过三个MedMNIST v2基准和四种架构研究了测试时间增强（TTA）对医学图像分类的影响，发现TTA在多数情况下会降低准确率，尤其在病理图像上降幅达31.6个百分点，但皮肤科图像中ResNet-18略有提升。
摘要：
测试时间增强（TTA）——通过多个增强版本的测试输入聚合预测——被广泛认为能提高分类准确性，特别是在医学影像中。本文通过三个MedMNIST v2基准和四种架构（参数量从21K到11M）进行系统实证研究，发现TTA在标准增强管道下相对于单次推理一致降低准确性，如ResNet-18在病理图像上的降幅达31.6个百分点。这种退化影响所有架构，包括卷积模型，并随着更多增强视图而恶化。唯一例外是ResNet-18在皮肤科图像上略有提升。我们发现增强输入与训练时间输入之间的分布偏移——放大了批量归一化统计信息不匹配——是主要机制。消融研究显示增强策略至关重要：仅强度增强比几何变换保留更多性能，包含原始未增强图像部分缓解但无法消除准确性下降。这些发现为实践者提供警示：TTA不应作为默认后置改进应用，而必须在特定模型-数据集组合上验证。
英文摘要：
Test-time augmentation (TTA)--aggregating predictions over multiple augmented copies of a test input--is widely assumed to improve classification accuracy, particularly in medical imaging where it is routinely deployed in production systems and competition solutions. We present a systematic empirical study challenging this assumption across three MedMNIST v2 benchmarks and four architectures spanning three orders of magnitude in parameter count (21K to 11M). Our principal finding is that TTA with standard augmentation pipelines consistently degrades accuracy relative to single-pass inference, with drops as severe as 31.6 percentage points for ResNet-18 on pathology images. This degradation affects all architectures, including convolutional models, and worsens with more augmented views. The sole exception is ResNet-18 on dermatology images, which gains a modest +1.6%. We identify the distribution shift between augmented and training-time inputs--amplified by batch normalization statistics mismatch--as the primary mechanism. Our ablation studies show that augmentation strategy matters critically: intensity-only augmentations preserve more performance than geometric transforms, and including the original unaugmented image partially mitigates but does not eliminate the accuracy drop. These findings serve as a cautionary note for practitioners: TTA should not be applied as a default post-hoc improvement but must be validated on the specific model-dataset combination.
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。
