| Presentation | 2018-05-14 Fundamental Research about Machine Classification of Twitter Images into Damage Types Munenari Inoguchi, Atsushi Imai, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Recently most people utilize smart-phone and post some information into social media such as Twitter and Facebook. Following this situation, much information relating to damage can be published by affected people in social medias. In those information, there could be many photos which represent actual indoor-damage and outdoor-damage. In order to retrieve damage information from photos posted in social media, we tried to utilize “Machine Classification” technology based on actual posted twitter data at 2016 Kumamoto Earthquake. First, we gathered twitter data tweeted in Kumamoto area with photos and geolocation, which amount was 5,853. Then, we retrieve 5,684 photos from those tweets, and we classified them into “indoor damage”, “building damage (outdoor damage)” and “road damage (outdoor damage)” as teacher data. The machine learned those teacher data in deep-learning method, and the machine tried to classify random-sampled photos into “indoor-damage” or “outdoor-damage”. We verified those result, and we found that over 80% photos can be classified adequately by machine classification without specified customization. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Machine Classification / Deep Learning / Damage Situation / Social Media |
| Paper # | ICTSSL2018-1,ASN2018-1 |
| Date of Issue | 2018-05-07 (ICTSSL, ASN) |
| Conference Information | |
| Committee | ASN / ICTSSL |
|---|---|
| Conference Date | 2018/5/14(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Hiroshima City Univ. |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Ambient intelligence, safe and secure life, poster session, etc. |
| Chair | Hiraku Okada(Nagoya Univ.) / Kazunori Okada(NICT) |
| Vice Chair | Shigeki Shiokawa(KAIT) / Jin Nakazawa(Keio Univ.) / Satoru Yamano(NEC) / Hiroshi Tamura(Chuo Univ.) / Keisuke Nakano(Niigata Univ.) |
| Secretary | Shigeki Shiokawa(NICT) / Jin Nakazawa(Sophia Univ.) / Satoru Yamano(NTT DoCoMo) / Hiroshi Tamura(Shizuoka Univ.) / Keisuke Nakano |
| Assistant | Hiroto Aida(Doshisha Univ.) / Tomoyuki Ota(Hiroshima City Univ.) / Tatsuya Kikuzuki(Fujitu Lab.) / Ryo Nakano(HITACHI) / Yoshifumi Hotta(Mitsubishi Electric) / Shosuke Sato(Tohoku Univ.) / Tomotaka Wada(Kansai Univ.) / Kazuyuki Miyakita(Niigata Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Ambient intelligence and Sensor Networks / Technical Committee on Information and Communication Technologies for Safe and Secure Life |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Fundamental Research about Machine Classification of Twitter Images into Damage Types |
| Sub Title (in English) | A Case Study of 2016 Kumamoto Earthquake |
| Keyword(1) | Machine Classification |
| Keyword(2) | Deep Learning |
| Keyword(3) | Damage Situation |
| Keyword(4) | Social Media |
| 1st Author's Name | Munenari Inoguchi |
| 1st Author's Affiliation | University of Toyama(Toyama Univ.) |
| 2nd Author's Name | Atsushi Imai |
| 2nd Author's Affiliation | NTT DATA CCS CORPORATION(NTT Data CCS) |
| Date | 2018-05-14 |
| Paper # | ICTSSL2018-1,ASN2018-1 |
| Volume (vol) | vol.118 |
| Number (no) | ICTSSL-26,ASN-27 |
| Page | pp.pp.1-4(ICTSSL), pp.1-4(ASN), |
| #Pages | 4 |
| Date of Issue | 2018-05-07 (ICTSSL, ASN) |