| 研究課題/領域番号 |
22KF0362
|
|---|
| 補助金の研究課題番号 |
21F21730 (2021-2022)
|
|---|
| 研究種目 |
特別研究員奨励費
|
| 配分区分 | 基金 (2023)
補助金 (2021-2022) |
|---|
| 応募区分 | 外国 |
|---|
| 審査区分 |
小区分44050:動物生理化学、生理学および行動学関連
|
|---|
| 研究機関 | 沖縄科学技術大学院大学 |
|---|
研究代表者 |
REITER Samuel 沖縄科学技術大学院大学, 計算行動神経科学ユ ニット, 准教授 (60869155)
|
|---|
| 研究分担者 |
ROUX NATACHA 沖縄科学技術大学院大学, 計算行動神経科学ユ ニット, 外国人特別研究員
|
|---|
| 研究期間 (年度) |
2023-03-08 – 2024-03-31
|
| 研究課題ステータス |
完了 (2023年度)
|
| 配分額 *注記 |
2,200千円 (直接経費: 2,200千円)
2023年度: 600千円 (直接経費: 600千円)
2022年度: 1,100千円 (直接経費: 1,100千円)
2021年度: 500千円 (直接経費: 500千円)
|
|---|
| キーワード | Cuttlefish / camouflage / natural images / statistics / marine biology / neuroscience / Visual perception / Camouflage / Cephalopod / Deep learning / Psychophysics |
|---|
| 研究開始時の研究の概要 |
By describing the image statistics used by camouflaging cuttlefish, we seek to understand how visual systems extract summaries of complex visual scenes in general.
|
| 研究実績の概要 |
Our work on this project has focused on 3 main topics. (1) Field Data Collection: We conducted diving expeditions, collecting hundreds of natural images from diverse habitats, forming a robust environmental database. Okinawan cuttlefish (Sepia latimanus) live in shallow water around the island, inhabiting at least 3 general habitats: coral reef, sand/coral rubble, and rocks. The typical camouflage pattern the animals adopt depends on the habitat, but also on the local image features in a particular habitat. (2) Camouflage Observation: We recorded images and video of cuttlefish camouflaging within these environments. This proved to be very technically difficult, owing to the high quality of the animals’ camouflage and to the skittish nature of these animals. (3) Statistical Analysis: Ongoing work is analyzing the statistical structure of the underwater natural images and quantifying the nature of camouflage pattern matching with these image statistics. The greatest challenge we face currently is in determining an appropriately high dimensional and accurate metric for camouflage pattern matching.
|
