Abstract
Detecting segments of interest from an input sequence is a challenging problem which often requires not only good knowledge of individual target segments, but also contextual understanding of the entire input sequence and the relationships between the target segments. To address this problem, we propose the Sequence-toSegments Network (), a novel end-to-end sequential encoder-decoder architecture. first encodes the input into a sequence of hidden states that progressively capture both local and holistic information. It then employs a novel decoding architecture, called Segment Detection Unit (SDU), that integrates the decoder state and encoder hidden states to detect segments sequentially. During training, we formulate the assignment of predicted segments to ground truth as the bipartite matching problem and use the Earth Mover’s Distance to calculate the localization errors. Experiments on temporal action proposal and video summarization show that achieves state-of-the-art performance on both tasks.