| Presentation | 2024-03-03 Assessment of the Utility of Tumor Location Information in MR Image Classification of Tumors Tsukasa Nishinakagawa, Yoshinari Takeishi, Jun'ichi Takeuchi, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | MRI, or magnetic resonance imaging, is a medical imaging technique widely used in various healthcare settings. It utilizes the magnetic resonance phenomenon of hydrogen nuclei within the body to obtain cross-sectional images of the body. In this research, we explore the classification problem of tumor types in head MR images using neural networks. Achieving high accuracy in such classifications is anticipated to contribute to automated diagnosis in clinical settings. However, a challenge in MR image classification is the limited availability of real-world data for training. To address this issue, our study employs fine-tuning with existing models to create an accurate model from a restricted dataset. Additionally, since the distribution of tumor occurrence varies based on tumor types, we investigate whether utilizing not only tumor shape but also its positional information can enhance the performance of MR image classification. We evaluate this by conducting experiments using publicly available MR images and their degraded versions. The results confirm that as images degrade, the utility of positional information increases. Notably, providing the positional information as an image proves more effective than presenting it as a 2D coordinate vector for enhancing the usefulness of positional information in the context of tumor classification. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | CNN / MRI / fine-tuning / image classification |
| Paper # | IBISML2023-43 |
| Date of Issue | 2024-02-25 (IBISML) |
| Conference Information | |
| Committee | PRMU / IBISML / IPSJ-CVIM |
|---|---|
| Conference Date | 2024/3/3(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Hiroshima Univ. Higashi-Hiroshima campus |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Kunio Kashio(NTT) / Masashi Sugiyama(Univ. of Tokyo) / 日浦 慎作(兵庫県立大) |
| Vice Chair | Takuya Funatomi(NAIST) / Go Irie(Tokyo Univ. of Science) / Toshihiro Kamishima(AIST) / Koji Tsuda(Univ. of Tokyo) |
| Secretary | Takuya Funatomi(Tokyo Inst. of Tech.) / Go Irie(Riken) / Toshihiro Kamishima(NTT) / Koji Tsuda(Hokkaido Univ.) / (名大) |
| Assistant | Kei Shimonishi(Kyoto Univ.) / Kensho Hara(AIST) / Yoshinobu Kawahara(Osaka Univ.) / Taiji Suzuki(Univ.of Tokyo) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding / Technical Committee on Information-Based Induction Sciences and Machine Learning / 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) | Assessment of the Utility of Tumor Location Information in MR Image Classification of Tumors |
| Sub Title (in English) | |
| Keyword(1) | CNN |
| Keyword(2) | MRI |
| Keyword(3) | fine-tuning |
| Keyword(4) | image classification |
| 1st Author's Name | Tsukasa Nishinakagawa |
| 1st Author's Affiliation | Kyushu University(Kyushu Univ.) |
| 2nd Author's Name | Yoshinari Takeishi |
| 2nd Author's Affiliation | Kyushu University(Kyushu Univ.) |
| 3rd Author's Name | Jun'ichi Takeuchi |
| 3rd Author's Affiliation | Kyushu University(Kyushu Univ.) |
| Date | 2024-03-03 |
| Paper # | IBISML2023-43 |
| Volume (vol) | vol.123 |
| Number (no) | IBISML-410 |
| Page | pp.pp.21-28(IBISML), |
| #Pages | 8 |
| Date of Issue | 2024-02-25 (IBISML) |