| Presentation | 2020-03-17 Object area extraction with biased training data set Ryo Nakamura, Yoshiaki Ueda, Jun Fujiki, Masaru Tanaka, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In order to extract an object region using machine learning, it is necessary to give a teacher label to each pixel. Providing a teacher label for each pixel is a very laborious task. In this paper, we propose a method to extract an area of an unidentified object without labeling every pixel by identifying the unidentified object with high precision. In this method, the training image set is not input as it is, but plural small patches are extracted from the training image set, and the discriminator is learned by using them as input. The learned discriminator extracts an object region by obtaining a discrimination ratio by using raster scan for an image to be extracted. At that time, though it is necessary to discriminate the non-discrimination object with high precision, by using the asymmetry of the training image set, the discrimination bias is utilized, and by discriminating the non-discrimination object with high precision, the region extraction of the discrimination object is carried out. And, visualization is carried out by making the identification ratio to be an image of αch. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | BTD / Object area extraction |
| Paper # | PRMU2019-88 |
| Date of Issue | 2020-03-09 (PRMU) |
| Conference Information | |
| Committee | PRMU / IPSJ-CVIM |
|---|---|
| Conference Date | 2020/3/16(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Yoichi Sato(Univ. of Tokyo) |
| Vice Chair | Toru Tamaki(Hiroshima Univ.) / Akisato Kimura(NTT) |
| Secretary | Toru Tamaki(NTT) / Akisato Kimura(OMRON SINICX) |
| Assistant | Yusuke Uchida(DeNA) / Takayoshi Yamashita(Chubu Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding / Special Interest Group on Computer Vision and Image Media |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Object area extraction with biased training data set |
| Sub Title (in English) | |
| Keyword(1) | BTD |
| Keyword(2) | Object area extraction |
| 1st Author's Name | Ryo Nakamura |
| 1st Author's Affiliation | Fukuoka University(Fukuoka Univ) |
| 2nd Author's Name | Yoshiaki Ueda |
| 2nd Author's Affiliation | Fukuoka University(Fukuoka Univ) |
| 3rd Author's Name | Jun Fujiki |
| 3rd Author's Affiliation | Fukuoka University(Fukuoka Univ) |
| 4th Author's Name | Masaru Tanaka |
| 4th Author's Affiliation | Fukuoka University(Fukuoka Univ) |
| Date | 2020-03-17 |
| Paper # | PRMU2019-88 |
| Volume (vol) | vol.119 |
| Number (no) | PRMU-481 |
| Page | pp.pp.115-119(PRMU), |
| #Pages | 5 |
| Date of Issue | 2020-03-09 (PRMU) |