| 研究課題/領域番号 |
16K16058
|
|---|
| 研究機関 | 国立情報学研究所 |
|---|
研究代表者 |
ュ イ 国立情報学研究所, コンテンツ科学研究系, 特任助教 (00754681)
|
|---|
| 研究期間 (年度) |
2016-04-01 – 2019-03-31
|
| キーワード | マルチモーダル分析 / 情報融合 / イベント検出 / パーソナライズド推薦 |
|---|
| 研究実績の概要 |
Ｗe presented the EventSensor system that aims to address sentics understanding and produces a multimedia summary for a given mood. It extracts concepts and mood tags from visual content and textual metadata of UGCs, and exploits them in supporting several significant multimedia-related services such as a musical multimedia summary. This work has been published in Journal of Knowledge-Based Systems.
We explored the fusion of multimodal information to refine tag ranking leveraging recall based weighting. This work has been published in ACM Multimedia COMMONS Workshop.
We proposed a tag recommendation system, called, PROMPT, that recommends personalized tags for a given photo leveraging personal and social contexts. This work has been published in IEEE International Symposium on Multimedia.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
1: 当初の計画以上に進展している
理由
The current research progress is going well. As I stated, we have published three works. We are working more tasks which are related to venue-related photo recognition and recommendation.
|
| 今後の研究の推進方策 |
We plan to investigate two related works: (i) introducing diversity in multimedia summaries by leveraging visual concepts in photos and (ii) enabling users to obtain a multimedia summary of any event and mood. Relevance and diversity are the two main characteristics of a good multimedia summary. We plan to address the diversity criterion in our systems by performing the clustering of photos during pre-processing. Clusters are formed based on concepts derived from the content of photos and helpful in producing diverse multimedia summaries. For instance, clustering based on visual concepts helps in producing a multimedia summary with visually dissimilar photos ( i.e. , from different clusters).
|
