All You Can Embed: Natural Language based Vehicle Retrieval with Spatio-Temporal Transformers

Combining Natural Language with Vision represents a unique and interesting\nchallenge in the domain of Artificial Intelligence. The AI City Challenge Track\n5 for Natural Language-Based Vehicle Retrieval focuses on the problem of\ncombining visual and textual information, applied to a smart-city use case. In\nthis paper, we present All You Can Embed (AYCE), a modular solution to\ncorrelate single-vehicle tracking sequences with natural language. The main\nbuilding blocks of the proposed architecture are (i) BERT to provide an\nembedding of the textual descriptions, (ii) a convolutional backbone along with\na Transformer model to embed the visual information. For the training of the\nretrieval model, a variation of the Triplet Margin Loss is proposed to learn a\ndistance measure between the visual and language embeddings. The code is\npublicly available at https://github.com/cscribano/AYCE_2021.\n