Faculty Profiles - TOKUNAGA Terumasa

Conference Prsentations (Oral, Poster) -

Division display1 - 20 of about 79／ All the affair displays >>

Cost-effective Deep Image Segmentation with Partial Patch Annotations based on Semi-supervised Learning

Terumasa Tokunaga, Keiichi Nakanishi, Ryoya Katafuchi, Kohki Miyama, Shou Watanabe

The 4th International Symposium on Neuromorphic AI Hardware 2022.12 Research Center for Neuromorphic AI Hardware

Event date： 2022.12.13 - 2022.12.14 Language：English Country：Japan

Image segmentation plays a central role in various real-world image processing applications, including object detection, object tracking, medical image diagnosis, scene understanding, and video surveillance. Over the last decade, supervised approaches based on deep learning (DL) have achieved outstanding performance in various segmentation tasks. However, most DL-based approaches require numerous pixel-wise annotations for training. This requirement often creates serious bottlenecks for research projects. Additionally, pixel-wise annotations of objects with indistinct boundaries can sometimes be arbitrary. In this study, we propose a novel semi-supervised image segmentation method called Cost-effective image Segmentation with Partial Annotations (CoSPA). Our main contributions are as follows: (1) Our method considerably reduces annotation costs of DL-based image segmentation maintaining practical performance, (2) We propose a novel cost function designed for CoSPA, (3) We experimentally demonstrate the effectiveness of CoSPA under extremely low annotation cost using real-world image datasets.

外観検査AIによる製造業のアップデート Invited

花手洋樹; 徳永旭将

エッセンスフォーラム2024 2024.08 株式会社エッセンス

画素単位のマスクを用いない大域・局所的推論の一致性に基づく領域分割

中西慶一, 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

Layer-wise External Attention for Efficient Deep Anomaly Detection

Tokihisa Hayakawa, Keiichi Nakanishi, Ryoya Katafuchi, Terumasa Tokunaga

3rd International Conference on Image Processing and Vision Engineering 2023.04

https://www.brain.kyutech.ac.jp/~neuro/2024/12/17/the-6th-international-symposium-on-neuromorphic-ai-hardware/?lang=en

Ren Takeyama, Keiichi Nakanishi, Terumasa Tokunaga

The 6th International Symposium on Neuromorphic AI Hardware 2025.03 the Society for Neuromorphic AI Hardware

Towards Resource-Efficient Machine Learning: Investigating Double Descent in Deep Mutual Learning

Fumika Komori, Keiichi Nakanishi, Terumasa Tokunaga

The 6th International Symposium on Neuromorphic AI Hardware 2025.03 the Society for Neuromorphic AI Hardware

Modeling Olfactory Adaptation in C. elegans Using Echo State Networks with Optimized Leak Rates

Kosuke Takashima, Ryosuke Ishibashi, Keiichi Nakanishi, Hakaru Tamukoh, Yuichiro Tanaka, Takeshi Ishihara, Noriko Sato, Terumasa Tokunaga

The 6th International Symposium on Neuromorphic AI Hardware 2025.03 the Society for Neuromorphic AI Hardware

Pooling-Enhanced Bidirectional 2D Reservoir Computing for Image Anomaly Detection

Keiichi Nakanishi, Ryosuke Ishibashi, Ren Takeyama, Terumasa Tokunaga

The 6th International Symposium on Neuromorphic AI Hardware 2025.03 the Society for Neuromorphic AI Hardware

Modeling Neural Responses in C. elegans Olfactory Neurons Using Deep Echo State Networks

Ryosuke Ishibashi, Keiichi Nakanishi, Yuichiro Tanaka, Hakaru Tamukoh, Takeshi Ishihara, Noriko Sato, Terumasa Tokunaga

The 6th International Symposium on Neuromorphic AI Hardware 2025.03 the Society for Neuromorphic AI Hardware

クラス事前確率を用いた半教師あり多クラス分類におけるコスト関数の提案

三山晃季, 中西慶一, 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

欠損補完に基づく画像異常検知に対してマスクサイズが与える影響の評価

古里和暉 , 中西慶一 , 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

外部視覚注視機構による異常検出性能の高い注視領域生成方法の解析

城間亮, 中西慶一, 片渕凌也, 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

EVSによる共振撮影を利用した透明物体のインスタンスセグメンテーションの検討

芦原新太郎, 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

欠損補完に基づく画像異常検知に対してマスクの色が与える影響の評価

持田海音, 中西慶一, 徳永旭将

第27回画像の認識・理解シンポジウム MIRU2024 2024.08 電子情報通信学会パターン認識・メディア理解(PRMU)研究専門委員会

外部視覚注視機構による深層学習の異常検出における注視領域生成過程の解析

城間亮、中西慶一、片渕凌也、徳永旭将

2024年度人工知能学会全国大会（第38回） 2024.05 人工知能学会

Evaluation of feature extractor trained by grayscale images for industrial anomaly detection

Sansei Hori, Kazumichi Tanaka, Keiichi Nakanishi, Terumasa Tokunaga

The 5th International Symposium on Neuromorphic AI Hardware 2024.03 the Society for Neuromorphic AI Hardware

Bidirectional 2D Reservoir Network for Image Anomaly Detection without any Training

Keiichi Nakanishi, Terumasa Tokunaga

The 5th International Symposium on Neuromorphic AI Hardware 2024.03 the Research Center for Neuromorphic AI Hardware, Kyushu Institute of Technology

Heavy Snow Cloud Detection in Satellite Images Based on Semi-Supervised Image Segmentation

Lin Magari, Terumasa Tokunaga, Kazue Suzuki

The 14th Symposium on Polar Science 2023.11 National Institute of Polar Research

The application of Echo State Network in reproducing the membrane potential response of olfactory neurons in C. elegans

IIBMP2023 2023.09

The Reproduction of Calcium Ion Response in C. elegans Olfactory Neurons using Echo State Networks

IIBMP2023 2023.09

　PREV- NEXT 　

<TOKUNAGA Terumasa>

Conference Prsentations (Oral, Poster) -