In most real-world industrial applications, such as multi-modal retrieval, traffic surveillance, video advertisement embedding, and automatic driving systems, information usually comes through multimedia data. For example, Web images in a multi-model retrieval system usually involve textual descriptions and multi-label tags; the videos in traffic surveillance contain both acoustic and visual signals; and sensory perceptions typically used in automatic driving system may need extensive multi-media data from multi-channel inputs in visual, auditory and motor pathways. Thereby, how to characterize the property of multimedia data so that it can be managed to enable different learning tasks of industrial applications is essential. This requires research to develop robust and sophisticated models to classify, retrieve and understand multi-media information. Moreover, with the explosion of multimedia data, the efficiency of model should also be considered so that systems can cope with the ever-demanding real-world industrial applications. In this special section, we look for cutting-edge techniques to efficiently handle multimedia data-information for sensing and computing in industrial applications.
Same submission guides as mentioned in the submission page apply. Please make sure to select the
correct special session from the dropdown menu