Abstract
Despite the success of deep learning in computer vision, algorithms to recognize subtle and small objects (or regions) is still challenging. For example, recognizing a baseball or a frisbee on a ground scene or a bone fracture in an X-ray image can easily result in overfitting, unless a huge amount of training data is available. To mitigate this problem, we need a way to force a model should identify subtle regions in limited training data. In this paper, we propose a simple but efficient supervised augmentation method called Cut\&Remain. It achieved better performance on various medical image domain (internally sourced- and public dataset) and a natural image domain (MS-COCO$_s$) than other supervised augmentation and the explicit guidance methods. In addition, using the class activation map, we identified that the Cut\&Remain methods drive a model to focus on relevant subtle and small regions efficiently. We also show that the performance monotonically increased along the Cut\&Remain ratio, indicating that a model can be improved even though only limited amount of Cut\&Remain is applied for, so that it allows low supervising (annotation) cost for improvement.
Abstract (translated)
URL
https://arxiv.org/abs/2111.13233
