Multimodal, Context-Aware, Feature Representation Learning for Classification and Localization | IEEE Conference Publication | IEEE Xplore