Fine-grained length controllable video captioning with ordinal embeddings
This work addresses the need for more flexible and precise caption generation in video understanding, offering incremental improvements over previous methods with limited length levels.
The paper tackles the problem of generating video captions with fine-grained length control by proposing two novel length embedding methods, bit and ordinal embeddings, which allow for precise control over caption length and timing. Experiments on ActivityNet Captions and Spoken Moments in Time datasets show effective length control, with analysis confirming separate learning of length and semantics.
This paper proposes a method for video captioning that controls the length of generated captions. Previous work on length control often had few levels for expressing length. In this study, we propose two methods of length embedding for fine-grained length control. A traditional embedding method is linear, using a one-hot vector and an embedding matrix. In this study, we propose methods that represent length in multi-hot vectors. One is bit embedding that expresses length in bit representation, and the other is ordinal embedding that uses the binary representation often used in ordinal regression. These length representations of multi-hot vectors are converted into length embedding by a nonlinear MLP. This method allows for not only the length control of caption sentences but also the control of the time when reading the caption. Experiments using ActivityNet Captions and Spoken Moments in Time show that the proposed method effectively controls the length of the generated captions. Analysis of the embedding vectors with ICA shows that length and semantics were learned separately, demonstrating the effectiveness of the proposed embedding methods.