Research

Our research interest focuses on Multimedia Analysis and Artificial Intelligence. Examples of our research topics include video retrieval, image recognition, and speech recognition. We are developing algorithms and systems on TSUBAME, TokyoTech's supercomputer. We welcome researchers and students who are interested in learning frameworks such as deep learning and zero-shot/few-shot learning.

井上研究室では、マルチメディア情報処理の分野で、映像検索・画像認識・音声認識などの人工知能技術に関する研究を実施しています。 大規模データを処理するためのアルゴリズムやシステムの開発では、東京工業大学のスーパーコンピューターTSUBAMEを活用しています。 深層学習やゼロショット学習などの学習方式や、人工知能を構築するためのデータ作成に興味のある研究者・学生を募集中です。

Members

 - Nakamasa Inoue (Associate Professor, contact e-mail: inoue at c.titech.ac.jp)
 - Undergraduate students: 2, Graduate students: 9

 - 井上 中順 (准教授, e-mail: inoue at c.titech.ac.jp)
 - 学部生 2名，大学院生 9名

Publications

• Zhibo Lou, Shinta Otake, Zhengxiao. Li, Rei Kawakami, Nakamasa Inoue, Cubic Knowledge Distillation for Speech Emotion Recognition, Proc. ICASSP, 2024.
• Sota Miyamoto, Takuma Yagi, Yuto Makimoto, Mahiro Ukai, Yoshitaka Ushiku, Atsushi Hashimoto, Nakamasa Inoue, PolarDB: Formula-Driven Dataset for Pre-Training Trajectory Encoders, Proc. ICASSP, 2024.
• Ryo Nakamura, Ryu Tadokoro, Eisuke Yamagata, Yusuke Kondo, Kensho Hara, Hirokatsu Kataoka, Nakamasa Inoue, Pseudo-outlier Synthesis Using q-Gaussian Distributions for Out-of-distribution Detection, Proc. ICASSP, 2024.
• Yanhao Bao, Tatsukichi Shibuya, Ikuro Sato, Rei Kawakami and Nakamasa Inoue, Efficient Target Propagation by Deriving Analytical Solution, Proc. AAAI, 2024.
• Taiki Miyanishi, Fumiya Kitamori, Shuhei Kurita, Jungdae Lee, Motoaki Kawanabe, Nakamasa Inoue CityRefer: Geography-aware 3D Visual Grounding Dataset on City-scale Point Cloud Data, Proc. NeurIPS (D&B track), 2023.
• Lei Xu, Rei Kawakami, Nakamasa Inoue, Scale-space Tokenization for Improving the Robustness of Vision Transformers, Proc. ACM Multimedia, 2023.
• Ryo Nakamura, Hirokatsu Kataoka, Sora Takashima, Edgar Josafat Martinez-Noriega, Rio Yokota, Nakamasa Inoue, Pre-training Vision Transformers with Very Limited Synthesized Images, Proc. ICCV, pp. 20360-20369, 2023.
• Risa Shinoda, Ryo Hayamizu, Kodai Nakashima, Nakamasa Inoue, Rio Yokota, Hirokatsu Kataoka, SegRCDB: Semantic Segmentation via Formula-Driven Supervised Learning, Proc. ICCV, pp. 20054-20063, 2023.
• Sora Takashima, Ryo Hayamizu, Nakamasa Inoue, Hirokatsu Kataoka, Rio Yokota, Visual Atoms: Pre-training Vision Transformers with Sinusoidal Waves, Proc. CVPR, pp. 18579-18588, 2023.
• Shinta Ohtake, Rei Kawakami, Nakamasa Inoue, Parameter Efficient Transfer Learning for Various Speech Processing Tasks, Proc. ICASSP, 2023.
• Keita Goto, Shinta Ohtake, Rei Kawakami, Nakamasa Inoue, Step Restriction for Improving Adversarial Attacks, Proc. ICASSP, 2023.
• Toshihiro Ota, Ikuro Sato, Rei Kawakami, Masayuki Tanaka, Nakamasa Inoue, Learning with Partial Forgetting in Modern Hopfield Networks, Proc. AISTATS, pp. 6661-6673, 2023.
• Tatsukichi Shibuya, Nakamasa Inoue, Rei Kawakami and Ikuro Sato, Fixed-Weight Difference Target Propagation, Proc. AAAI, pp. 9811-9819, 2023.
• Ruoyue Shen, Nakamasa Inoue, and Koichi Shinoda, Text-Guided Object Detector for Multi-modal Video Question Answering, Proc. WACV, pp. 1032-1042, 2023.
• Ikuro Sato, Ryota Yamada, Masayuki Tanaka, Nakamasa Inoue, Rei Kawakami, PoF: Post-Training of Feature Extractor for Improving Generalization, Proc. ICML, pp. 19221-19230, 2022.
• Hirokatsu Kataoka, Ryo Hayamizu, Ryosuke Yamada, Kodai Nakashima, Sora Takashima, Xinyu Zhang, Edgar Josafat Martinez-Noriega, Nakamasa Inoue, Rio Yokota, Replacing Labeled Real-image Datasets with Auto-generated Contours, Proc. CVPR, pp. 21232-21241, 2022.
• Tomohiro Hayase, Suguru Yasutomi, Nakamasa Inoue, Downstream Augmentation Generation for Contrastive Learning, Proc. ICASSP, pp. 2115-2119, 2022.
• Hirokatsu Kataoka, Kazushige Okayasu, Asato Matsumoto, Eisuke Yamagata, Ryosuke Yamada, Nakamasa Inoue, Akio Nakamura, Yutaka Satoh, Pre-training without Natural Images, IJCV, vol. 130, pp. 990-1007, 2022.
• Kodai Nakashima, Hirokatsu Kataoka, Asato Matsumoto, Kenji Iwata, Nakamasa Inoue, Yutaka Satoh, Can Vision Transformers Learn without Natural Images?, Proc. AAAI, pp. 1990-1998, 2022.
• Hirokatsu Kataoka, Kensho Hara, Ryusuke Hayashi, Eisuke Yamagata, Nakamasa Inoue, Spatiotemporal Initialization for 3D CNNs with Generated Motion Patterns, Proc. WACV, pp. 1279-1288, 2022.
• Nakamasa Inoue, Tsubasa Maruyama, Keita Goto, Augmentation-Agnostic Regularization for Unsupervised Contrastive Learning with Its Application to Speaker Verification Proc. APSIPA, pp. 1993-1998, 2021. (Best Paper Award)
• Hirokatsu Kataoka, Asato Matsumoto, Eisuke Yamagata, Ryosuke Yamada, Nakamasa Inoue, Yutaka Satoh, Formula-Driven Supervised Learning with Recursive Tiling Patterns, Proc. ICCV HTCV Workshop, pp. 4098-4105, 2021.
• Keita Goto, Nakamasa Inoue, Learning VAE with Categorical Labels for Generating Conditional Handwritten Characters, Proc. MVA, 2021.
• Mariana Rodrigues Makiuchi, Tifani Warnita, Nakamasa Inoue, Koichi Shinoda, Michitaka Yoshimura, Momoko Kitazawa, Kei Funaki, Yoko Eguchi, Taishiro Kishimoto, Speech Paralinguistic Approach for Detecting Dementia Using Gated Convolutional Neural Network, IEICE Trans. on Information and Systems, vol. E104–D, no.11, pp. 1930-1940, 2021.
• Ryota Nakazawa, Yuki Minamoto, Nakamasa Inoue, and Mamoru Tanahashi, Species Reaction Rate Modelling based on Physics-Guided Machine Learning, Elsevier Combustion and Flame, vol. 235, pp. 1-11, 2021.
• Nakamasa Inoue, Ryota Yamada, Rei Kawakami, Ikuro Sato, Disentangling Latent Groups of Factors, Proc. ICIP, 2021.
• Nakamasa Inoue, Teacher-Assisted Mini-Batch Sampling for Blind Distillation using Metric Learning, Proc. ICASSP, 2021.
• Hikaru Nakata, Nakamasa Inoue, Rio Yokota, Self-supervised Continual Pretraining for Class Incremental Image Classification, Proc. CVPR CLVISION Workshop, 2021.
• Kherlen Jigjid, Chitoshi Tamaoki, Yuki Minamoto, Ryota Nakazawa, Nakamasa Inoue, Mamoru Tanahashi. Data Driven Analysis and Prediction of MILD Combustion Mode, Elsevier Combustion and Flame, vol. 223, pp. 475-485, 2021.
• Hirokatsu Kataoka, Kazushige Okayasu, Asato Matsumoto, Eisuke Yamagata, Ryosuke Yamada, Nakamasa Inoue, Akio Nakamura, Yutaka Satoh, Pre-training without Natural Images, Proc. ACCV, 2020. (Best Paper Honorable Mention Award)
• Nakamasa Inoue, Graph Grouping Loss for Metric Learning of Face Image Representations, Proc. VCIP, 2020.
• Nakamasa Inoue, Keita Goto, Semi-Supervised Contrastive Learning with Generalized Contrastive Loss and Its Application to Speaker Recognition, Proc. APSIPA, 2020.
• Keita Goto, Nakamasa Inoue, Quasi-Newton Adversarial Attacks on Speaker Verification Systems, Proc. APSIPA, 2020.
• Nakamasa Inoue, Keita Goto, Optimizing Speaker Embeddings using Meta-Training Sets, Proc. APSIPA, 2020.
• Nakamasa Inoue, Keita Goto, Closed-Form Pre-Training for Small-Sample Environmental Sound Recognition, Proc. APSIPA, 2020.
• Takehiko Ohkawa, Naoto Inoue, Hirokatsu Kataoka, Nakamasa Inoue, Augmented Cyclic Consistency Regularization for Unpaired Image-To-Image Translation, Proc. ICPR, 2020.
• Nakamasa Inoue, Eisuke Yamagata, Hirokatsu Kataoka. Initialization Using Perlin Noise for Training Networks with a Limited Amount of Data, Proc. ICPR, 2020.
• Raden Mu'az Mun'im, Nakamasa Inoue, Koichi Shinoda, Sequence-Level Knowledge Distillation for Model Compression of Attention-Based Sequence-to-Sequence Speech Recognition Proc. ICASSP, pp. 6151-6155, 2019.
• Nakamasa Inoue, Koichi Shinoda, Few-Shot Adaptation for Multimedia Semantic Indexing, Proc. ACM Multimedia, pp. 1110-1118, 2018.
• Jiacen Zhang, Nakamasa Inoue, Koichi Shinoda, I-vector Transformation Using Conditional Generative Adversarial Networks for Short Utterance Speaker Verification, Proc. Interspeech, pp. 3613-3617, 2018.
• Tifani Warnita, Nakamasa Inoue, Koichi Shinoda, Detecting Alzheimer's Disease Using Gated Convolutional Neural Network from Audio Data, Proc. Interspeech, pp. 1706-1710, 2018.
• Thao Minh Le, Nakamasa Inoue, Koichi Shinoda, A Fine-to-Coarse Convolutional Neural Network for 3D Human Action Recognition, Proc. British Machine Vision Conference (BMVC), 2018.
• Haoyi Zhang, Conggui Liu, Nakamasa Inoue, Koichi Shinoda, Multi-Task Autoencoder for Noise-Robust Speech Recognition, Proc. ICASSP, pp. 5599-5603, 2018.
• Tommi Kerola, Nakamasa Inoue and Koichi Shinoda, Cross-View Human Action Recognition from Depth Maps Using Spectral Graph Sequences, Elsevier Journal of Computer Vision and Image Understanding, vol. 154, pp. 108–126, 2017.
• Yuki Yasui, Nakamasa Inoue, Koji Iwano, and Koichi Shinoda, Multimodal Speech Recognition Using Mouth Images from Depth Camera, Proc. APSIPA, pp. 128–132, 2017.
• Conggui Liu, Nakamasa Inoue, and Koichi Shinoda, A Unified Network for Multi-Speaker Speech Recognition with Multi-Channel Recordings, Proc. APSIPA, pp. 160–165, 2017.
• Mengxi Lin, Nakamasa Inoue, and Koichi Shinoda, CTC Network with Statistical Language Modeling for Action Sequence Recognition in Videos, Proc. ACM Multimedia Thematic Workshop, 2017.
• Shinya Matsui, Nakamasa Inoue, Yuko Akagi, Goshu Nagino, and Koichi Shinoda, User Adaptation of Convolutional Neural Network for Human Activity Recognition, Proc. EUSIPCO, pp. 753–757, 2017.
• Nakamasa Inoue and Koichi Shinoda, Semantic Indexing for Large-Scale Video Retrieval (Invited Paper), ITE Trans. on Media Technology and Applications, vol. 4, no. 3, pp. 209–217, 2016.
• Tommi Kerola, Nakamasa Inoue and Koichi Shinoda, Graph Regularized Implicit Pose for 3D Human Action Recognition, Proc. APSIPA, pp. 155–159, 2016.
• Nakamasa Inoue and Koichi Shinoda, Adaptation of Word Vectors using Tree Structure for Visual Semantics, Proc. ACM Multimedia, 2016.
• Nakamasa Inoue and Koichi Shinoda, Fast Coding of Feature Vectors using Neighbor-To-Neighbor Search, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 38, no. 6, pp. 1170–1184, 2015.
• Nakamasa Inoue and Koichi Shinoda, Vocabulary Expansion Using Word Vectors for Video Semantic Indexing, Proc. ACM Multimedia, 2015.
• Fumito Nishi, Nakamasa Inoue and Koichi Shinoda, Speaker Diarization Using Multi-Modal i-vectors, Proc. ITC-CSCC, 2015.
• Nakamasa Inoue and Koichi Shinoda, n-Gram Models for Video Semantic Indexing, Proc. ACM Multimedia, 2014.
• Tommi Kerola, Nakamasa Inoue, and Koichi Shinoda, Spectral Graph Skeletons for 3D Action Recognition, Proc. ACCV, 2014.
• Zhuolin Liang, Nakamasa Inoue and Koichi Shinoda, Event Detection by Velocity Pyramid, Proc. Multimedia Modeling, 2014.
• Nakamasa Inoue and Koichi Shinoda, q-Gaussian Mixture Models for Image and Video Semantic Indexing, Elsevier Journal of Visual Communication and Image Representation, vol. 24, no. 8, pp. 1450–1457, 2013.
• Yusuke Kamishima, Nakamasa Inoue and Koichi Shinoda, Event Detection in Con- sumer Videos Using GMM Supervectors and SVMs, EURASIP Journal on Image and Video Processing, vol. 51, pp. 1–13, 2013.
• Nakamasa Inoue and Koichi Shinoda, Neighbor-To-Neighbor Search for Fast Coding of Feature Vectors, Proc. ICCV, 2013.
• Nakamasa Inoue and Koichi Shinoda, q-Gaussian Mixture Models Based on Non- Extensive Statistics for Image and Video Semantic Indexing, Proc. ACCV, 2012.
• Nakamasa Inoue and Koichi Shinoda, A Fast and Accurate Video Semantic-Indexing System Using Fast MAP Adaptation and GMM Supervectors, IEEE Transactions on Multimedia, vol. 14, no. 4, pp. 1196–1205, 2012.
• Yusuke Kamishima, Nakamasa Inoue, Koichi Shinoda, and Shunsuke Sato, Multimedia Event Detection Using GMM Supervectors and SVMs, Proc. ICIP, 2012.
• Nakamasa Inoue and Koichi Shinoda, A Fast MAP Adaptation Technique for GMM- supervector-based Video Semantic Indexing, Proc. ACM Multimedia, 2011.
• Nakamasa Inoue, Tatsuhiko Saito, Koichi Shinoda, and Sadaoki Furui, High-Level Feature Extraction using SIFT GMMs and Audio Models, Proc. ICPR, 2010.

• 澁谷 辰吉, 井上 中順, 川上 玲, 佐藤 育郎, 二値重み空間でのBinary Neural Networksの学習, 画像の認識・理解シンポジウム MIRU, 2023. (MIRUインタラクティブ発表賞受賞)
• 篠田 理沙, 速水 亮, 中嶋 航大, 井上 中順, 横田 理央, 片岡 裕雄, 数式ドリブン教師あり学習によるセマンティックセグメンテーション, 画像の認識・理解シンポジウム MIRU, 2023. (MIRU学生奨励賞受賞)
• Ryosuke Yamada, Kensho Hara, Hirokatsu Kataoka, Koshi Makihara, Nakamasa Inoue, Rio Yokota, Yutaka Satoh Formula-Supervised Visual-Geometric Pre-training, 画像の認識・理解シンポジウム MIRU, 2023. (MIRU学生奨励賞受賞)
• Kenji Cari Koga, Nakamasa Inoue, Rei Kawakami, Trajectory Collection with Few-Shot Imitation Learning and Proximal Policy Optimization, 画像の認識・理解シンポジウム MIRU, 2023.
• 佐藤 郎真, 井上 中順, 川上 玲, ロボットによる物体再配置における多変量正規分布を用いた衝突回避 画像の認識・理解シンポジウム MIRU, 2023.
• 髙橋 秀弥, 井上 中順, 横田 理央, 片岡 裕雄, 前田 英作, 学習過程における形状・テクスチャ偏重度の推移と事前学習データセットとの関係について, 画像の認識・理解シンポジウム MIRU, 2023.
• 速水 亮, 高島 空良, 井上 中順, 片岡 裕雄, 横田 理央, Visual Atoms: 正弦波の輪郭表現によるVision Transformerの事前学習, 画像センシングシンポジウム SSII, 2023. (SSIIオーディエンス賞受賞)
• 中村 凌, 片岡 裕雄, 高島 空良, Martinez-Noriega Edgar Josafat, 横田 理央, 井上 中順, 限られた合成画像を用いたVision Transformerの事前学習, 画像センシングシンポジウム SSII, 2023. (SSIIオーディエンス賞受賞)
• 浅倉 拓也, 井上 中順, 横田 理央, 篠田 浩一, 受容野の自動最適化によるモードに適応的なTransformerの開発, 人工知能学会 JSAI, 2023.
• 髙橋 秀弥, 井上 中順, 横田 理央, 片岡 裕雄, 前田 英作, 画像識別における形状・テクスチャ偏重度と二重降下現象の関係について, パターン認識・メディア理解研究会 PRMU, 2023.
• Jungdae Lee, Rei Kawakami, Nakamasa Inoue, Similarity Based Attention on a Hypersphere for Vision Transformers, パターン認識・メディア理解研究会 PRMU, 2022.
• 杉山 佳史, 片岡 裕雄, 横田 理央, 井上 中順, 敵対的距離学習モジュールを用いた特徴変動に頑健な画像認識のための対照学習, パターン認識・メディア理解研究会 PRMU, 2022.
• 篠田 理沙, 速水 亮, 中嶋 航大, 井上 中順, 片岡 裕雄, セグメンテーションタスクにおける実画像を用いない事前学習, ビジョン技術の実利用ワークショップ ViEW, 2022. (若手奨励賞受賞)
• 田所 龍, 片岡 裕雄, 川上 玲, 横田 理央, 井上 中順, 蒸留画像による事前学習効果についての検討, ビジョン技術の実利用ワークショップ ViEW, 2022.
• 宮本 蒼太, 八木 拓真, 牛久 祥孝, 橋本 敦史, 井上 中順, 手の軌道特徴を用いた一人称視点料理動画における詳細動作認識, パターン認識・メディア理解研究会 PRMU, 2022. (PRMU月間ベストプレゼンテーション賞受賞)
• 澁谷 辰吉, 佐藤 育郎, 川上 玲, 井上 中順, Random Matrixによる固定逆伝播ネットワークを用いたTarget Propagation派生手法の提案, 画像の認識・理解シンポジウム MIRU, 2022. (MIRUインタラクティブ発表賞受賞)
• 山田 陵太, 佐藤 育郎, 田中 正行, 井上 中順, 川上 玲, 深層モデルの汎化性能改善を目的とした特徴抽出器の事後学習, 画像の認識・理解シンポジウム MIRU, 2022. (MIRU長尾賞受賞, テレコムシステム技術学生賞受賞)
• Toshihiro Ota, Ikuro Sato, Rei Kawakami, Masahiro Tanaka, Nakamasa Inoue, Learning with Partial Forgetting in Modern Hopfield Networks, 画像の認識・理解シンポジウム MIRU, 2022.
• 江良 真結子, 井上 中順, 篠田 浩一, 細川 稜平, 村田 勝寛, 庭野 聖史, 谷津 陽一, 河合 誠之, ロバスト主成分分析に基づく劣化CCDカメラ画像のノイズ除去, 日本天文学会春季年会, 2022.
• 伊藤 尚泰, 村田 勝寛, 細川 稜平, 笹田 真人, 庭野 聖史, 谷津 陽一, 河合 誠之, 篠田 浩 一, 井上 中順, 伊藤 亮介, 下川辺 隆史, MITSuME望遠鏡画像に対する深層学習を用いた突発天体検知システムの構築, 日本天文学会春季年会, 2022.
• 松本 晨人, 山縣 英介, 井上 中順, 片岡 裕雄, 佐藤 雄隆, 数式ベースの事前学習用データセットの特性評価, 画像センシングシンポジウム SSII, 2020.
• Kodai Nakashima, Hirokatsu Kataoka, Asato Matsumoto, Kenji Iwata, Nakamasa Inoue, Can Vision Transformers Learn without Natural Images?, 画像の認識・理解シンポジウム MIRU, 2021.
• Takehiko Ohkawa, Naoto Inoue, Hirokatsu Kataoka, Nakamasa Inoue, Consistency Regularization using Data Augmentation for Cycle-Consistent GANs, 画像の認識・理解シンポジウム MIRU, 2021.
• 井上 中順, 限られたデータからの深層学習 (チュートリアル講演), 画像の認識・理解シンポジウム MIRU, 2021.
• 井上 中順, 限られたデータからの学習法 (招待講演), 精密工学会 定例研究会, 2020.
• Jigjid Kherlen, Yuki Minamoto, Nakazawa Ryota, Inoue Nakamasa, Tanahashi Mamoru, A DNN based identifier for MILD combustion mode in an LES context, 日本伝熱学会 日本伝熱シンポジウム, 2020.
• 井上 中順, 限られたデータからの深層学習 (オーガナイズドセッション), 画像センシングシンポジウム SSII, 2020.
• Kengo Machida, Nakamasa Inoue, Koichi Shinoda. KL統計量に基づくニューラルネットワークのプルーニング, 画像の認識・理解シンポジウム MIRU, 2019.

Access

G3-1013 (Mail-Box:G3-55), Tokyo Institute of Technology (Institute of Science Tokyo), 4259 Nagatsuta, Midori-ku, Yokohama, 226-8502, Japan.

〒226-8502 横浜市緑区長津田町4259 東京工業大学 G3棟1013号室(ポスト:G3-55) ※ 2024年10月より東京科学大学 (科学大)