On the Importance of Visual Context for Data Augmentation in Scene Understanding
This work addresses the challenge of effective data augmentation for computer vision tasks, particularly in scenarios with limited annotations, but it is incremental as it builds on existing augmentation methods by adding context modeling.
The authors tackled the problem of data augmentation for scene understanding tasks by proposing a context model to place objects appropriately in images, which improved object detection, semantic and instance segmentation on PASCAL VOC12 and COCO datasets, with significant gains in limited annotation scenarios.
Performing data augmentation for learning deep neural networks is known to be important for training visual recognition systems. By artificially increasing the number of training examples, it helps reducing overfitting and improves generalization. While simple image transformations can already improve predictive performance in most vision tasks, larger gains can be obtained by leveraging task-specific prior knowledge. In this work, we consider object detection, semantic and instance segmentation and augment the training images by blending objects in existing scenes, using instance segmentation annotations. We observe that randomly pasting objects on images hurts the performance, unless the object is placed in the right context. To resolve this issue, we propose an explicit context model by using a convolutional neural network, which predicts whether an image region is suitable for placing a given object or not. In our experiments, we show that our approach is able to improve object detection, semantic and instance segmentation on the PASCAL VOC12 and COCO datasets, with significant gains in a limited annotation scenario, i.e. when only one category is annotated. We also show that the method is not limited to datasets that come with expensive pixel-wise instance annotations and can be used when only bounding boxes are available, by employing weakly-supervised learning for instance masks approximation.