| Presentation | 2020-01-29 A study of generalized generation of image features for computer-aided detection systems based on unsupervised learning with normal datasets Kazuyuki Ushifusa, Mitsutaka Nemoto(, Yuichi Kimura, Takashi Nagaoka, Takahiro Yamada, Atsuko Tanaka, Naoto Hayashi, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In a computer-aided detection system, image features are essential factors. In this study, we propose an image feature generation method that is based on unsupervised deep learning with only a normal dataset and could generate image features irrespective of the training dataset scale. To evaluate robustness against the scale of training data, we experimentally evaluate change of performance with the reduction of the scale of the training dataset. As a result of applied the proposed method to the identification of cerebral aneurysm on head MRA, the average ANODE score was 0.523 ± 0.0362. Furthermore, we also confirmed that our method could create useful features, even if the training data decrease. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Image feature / Unsupervised deep learning / Convolutional autoencoder / Small training dataset |
| Paper # | MI2019-68 |
| Date of Issue | 2020-01-22 (MI) |
| Conference Information | |
| Committee | MI |
|---|---|
| Conference Date | 2020/1/29(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | OKINAWAKEN SEINENKAIKAN |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Medical Image Engineering, Analysis, Recognition, etc. |
| Chair | Yoshiki Kawata(Tokushima Univ.) |
| Vice Chair | Takayuki Kitasaka(Aichi Inst. of Tech.) / Hidekata Hontani(Nagoya Inst. of Tech.) |
| Secretary | Takayuki Kitasaka(Yamaguchi Univ.) / Hidekata Hontani(Univ. of Hyogo) |
| Assistant | Hotaka Takizawa(Tsukuba Univ.) / Yoshito Otake(NAIST) |
| Paper Information | |
| Registration To | Technical Committee on Medical Imaging |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A study of generalized generation of image features for computer-aided detection systems based on unsupervised learning with normal datasets |
| Sub Title (in English) | Experimental evaluations of feature generation by small datasets |
| Keyword(1) | Image feature |
| Keyword(2) | Unsupervised deep learning |
| Keyword(3) | Convolutional autoencoder |
| Keyword(4) | Small training dataset |
| 1st Author's Name | Kazuyuki Ushifusa |
| 1st Author's Affiliation | Kindai University(Kindai Uni.) |
| 2nd Author's Name | Mitsutaka Nemoto( |
| 2nd Author's Affiliation | Kindai University(Kindai Uni.) |
| 3rd Author's Name | Yuichi Kimura |
| 3rd Author's Affiliation | Kindai University(Kindai Uni.) |
| 4th Author's Name | Takashi Nagaoka |
| 4th Author's Affiliation | Kindai University(Kindai Uni.) |
| 5th Author's Name | Takahiro Yamada |
| 5th Author's Affiliation | Kindai University(Kindai Uni.) |
| 6th Author's Name | Atsuko Tanaka |
| 6th Author's Affiliation | Kindai University(Kindai Uni.) |
| 7th Author's Name | Naoto Hayashi |
| 7th Author's Affiliation | The University of Tokyo Hospital(The Uni of Tokyo Hosp) |
| Date | 2020-01-29 |
| Paper # | MI2019-68 |
| Volume (vol) | vol.119 |
| Number (no) | MI-399 |
| Page | pp.pp.15-18(MI), |
| #Pages | 4 |
| Date of Issue | 2020-01-22 (MI) |