| Project/Area Number |
20K19822
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 61010:Perceptual information processing-related
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
Garcia Docampo Noa 大阪大学, データビリティフロンティア機構, 特任助教(常勤) (80870005)
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Project Status |
Completed (Fiscal Year 2022)
|
| Budget Amount *help |
¥2,860,000 (Direct Cost: ¥2,200,000、Indirect Cost: ¥660,000)
Fiscal Year 2021: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2020: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
|
|---|
| Keywords | computer vision for art / vision and language / description generation / image search / computer vision / art analysis / text-to-image generation / visual representation / digital humanities / automatic art analysis / art description / artwork recognition / visual recognition / image captioning / visual understanding / deep neural networks / transfer learning / question answering |
|---|
| Outline of Research at the Start |
This research studies how machines can understand and interpret art as we humans do. When we look at a painting, we "read" it by recognizing its depicted objects, its emotions, its style, etc. By using machine learning tools, we will replicate this behavior and provide new interpretations of art.
|
| Outline of Final Research Achievements |
In this project, we have conducted research to understand concret and abstract representation in art using computer vision (CV) and artificial intelligence (AI) techniques. The achievements of the project are three fold:
1) We created a dataset for answering art-related questions about fine-art paintings. 2) We developed a model to generate descriptions from paintings. 3) We created a dataset for identifying artworks given photo.
Our work was published at top conferences and delivered to the research community through several invited talks.
|
| Academic Significance and Societal Importance of the Research Achievements |
With this research, we aimed to make art more accessible to people by developing artifical intelligence models that facilitate the understanding of art. For example, by generating automatic explanations of paintings, museum visitors can understand the intrinsic meaning of each artwork easily.
|
