Self corrective Perturbations for Semantic Segmentation and Classification
This addresses the challenge of boosting model accuracy post-training for computer vision practitioners, though it appears incremental as it builds on existing methods.
The paper tackles the problem of improving pre-trained CNNs without retraining by introducing guided perturbations to inputs, resulting in enhanced prediction performance on semantic segmentation and classification tasks using datasets like PASCAL VOC, MNIST, and CIFAR10.
Convolutional Neural Networks have been a subject of great importance over the past decade and great strides have been made in their utility for producing state of the art performance in many computer vision problems. However, the behavior of deep networks is yet to be fully understood and is still an active area of research. In this work, we present an intriguing behavior: pre-trained CNNs can be made to improve their predictions by structurally perturbing the input. We observe that these perturbations - referred as Guided Perturbations - enable a trained network to improve its prediction performance without any learning or change in network weights. We perform various ablative experiments to understand how these perturbations affect the local context and feature representations. Furthermore, we demonstrate that this idea can improve performance of several existing approaches on semantic segmentation and scene labeling tasks on the PASCAL VOC dataset and supervised classification tasks on MNIST and CIFAR10 datasets.