End-to-end Concept Word Detection for Video Captioning, Retrieval, and Question Answering

End-to-end Concept Word Detection for Video Captioning, Retrieval, and Question Answering

Youngjae Yu,  Hyungjin Ko,  Jongwook Choi,  Gunhee Kim

(Submitted on 10 Oct 2016 ( v1), last revised 13 Dec 2016 (this version, v2))

We propose a high-level concept word detector that can be integrated with any video-to-language models. It takes a video as input and generates a list of concept words as useful semantic priors for language generation models. The proposed word detector has two important properties. First, it does not require any external knowledge sources for training. Second, the proposed word detector is trainable in an end-to-end manner jointly with any video-to-language models. To maximize the values of detected words, we also develop a semantic attention mechanism that selectively focuses on the detected concept words and fuse them with the word encoding and decoding in the language model. In order to demonstrate that the proposed approach indeed improves the performance of multiple video-to-language tasks, we participate in four tasks of LSMDC 2016. Our approach achieves the best accuracies in three of them, including fill-in-the-blank, multiple-choice test, and movie retrieval. We also attain comparable performance for the other task, movie description.

Comments:	20 pages, 12 figures; Winner of three (fill-in-the-blank, multiple-choice test, and movie retrieval) out of four tasks of the LSMDC 2016 Challenge (Workshop in ECCV 2016)
Subjects:	Computer Vision and Pattern Recognition (cs.CV)
Cite as:	arXiv:1610.02947 [cs.CV]
	(or arXiv:1610.02947v2 [cs.CV] for this version)

Submission history

From: Jongwook Choi [ view email] 
[v1] Mon, 10 Oct 2016 15:03:15 GMT (2876kb,D)
[v2] Tue, 13 Dec 2016 14:27:20 GMT (3934kb,D)