Return of the Devil in the Details: Delving Deep into Convolutional Nets
This work provides a comparative analysis for researchers in computer vision, clarifying implementation details and performance trade-offs between deep and shallow methods, though it is incremental in nature.
The paper conducted a rigorous evaluation of Convolutional Neural Networks (CNNs) compared to shallow representations like Bag-of-Visual-Words and Improved Fisher Vector, identifying that CNN output dimensionality can be reduced significantly without performance loss and that data augmentation techniques benefit both deep and shallow methods.
The latest generation of Convolutional Neural Networks (CNN) have achieved impressive results in challenging benchmarks on image recognition and object detection, significantly raising the interest of the community in these methods. Nevertheless, it is still unclear how different CNN methods compare with each other and with previous state-of-the-art shallow representations such as the Bag-of-Visual-Words and the Improved Fisher Vector. This paper conducts a rigorous evaluation of these new techniques, exploring different deep architectures and comparing them on a common ground, identifying and disclosing important implementation details. We identify several useful properties of CNN-based representations, including the fact that the dimensionality of the CNN output layer can be reduced significantly without having an adverse effect on performance. We also identify aspects of deep and shallow methods that can be successfully shared. In particular, we show that the data augmentation techniques commonly applied to CNN-based methods can also be applied to shallow methods, and result in an analogous performance boost. Source code and models to reproduce the experiments in the paper is made publicly available.