Timezone: »

Learning Loss for Test-Time Augmentation
Ildoo Kim · Younghoon Kim · Sungwoong Kim

Thu Dec 10 09:00 PM -- 11:00 PM (PST) @ Poster Session 6 #1883

Data augmentation has been actively studied for robust neural networks. Most of the recent data augmentation methods focus on augmenting datasets during the training phase. At the testing phase, simple transformations are still widely used for test-time augmentation. This paper proposes a novel instance-level test- time augmentation that efficiently selects suitable transformations for a test input. Our proposed method involves an auxiliary module to predict the loss of each possible transformation given the input. Then, the transformations having lower predicted losses are applied to the input. The network obtains the results by averaging the prediction results of augmented inputs. Experimental results on several image classification benchmarks show that the proposed instance-aware test- time augmentation improves the model’s robustness against various corruptions.

Author Information

Ildoo Kim (Kakao Brain)
Younghoon Kim (Sungshin Women's University)
Sungwoong Kim (Kakao Brain)

More from the Same Authors