Dense Image Representation with Spatial Pyramid VLAD Coding of CNN for Locally Robust Captioning
This is an incremental improvement for image captioning systems, addressing the issue of overlooking secondary elements in images.
The paper tackled the problem of CNN features in image captioning often missing local image elements by proposing a VLAD coding method on spatial pyramid for sub-region features, achieving captions that better account for local details with only 3% of the dimensionality of CNN features.
The workflow of extracting features from images using convolutional neural networks (CNN) and generating captions with recurrent neural networks (RNN) has become a de-facto standard for image captioning task. However, since CNN features are originally designed for classification task, it is mostly concerned with the main conspicuous element of the image, and often fails to correctly convey information on local, secondary elements. We propose to incorporate coding with vector of locally aggregated descriptors (VLAD) on spatial pyramid for CNN features of sub-regions in order to generate image representations that better reflect the local information of the images. Our results show that our method of compact VLAD coding can match CNN features with as little as 3% of dimensionality and, when combined with spatial pyramid, it results in image captions that more accurately take local elements into account.