Takumi Iida

Takumi Iida

Machine Learning/ Computer Vision

Japanese page has more information!

Skills

  • Computer Vision 60%
  • Computer Graphics 40%
  • LLM 20%

Education

  • Master of Electronic System Ritsumeikan University 2018-2020
  • Bachelor of Electronic and Computer Engineering Ritsumeikan University 2014-2018

Experience

LayerX Inc.

Machine Learning Engineer

2025.3 - Present Remote

Elith Inc.

Research Manager

2024.10 - 2025.2 Remote
Research and development of multi-agent systems for LLM content inference

Developed a multi-agent system to infer content from LLM titles. Aimed for co-authoring a paper, and the project was accepted at ICLR 2025 Workshop and JSAI 2025.

LPIXEL Inc.

Machine Learning Engineer

2024.4 - 2024.9 Remote
Anomaly Detection in X-ray Images

Developed a model to detect anomalies from chest X-ray images.

SenseTime Japan, Ltd.

Researcher

2020.4 - 2023.3 Kyoto
Fallen Object Detection Project

Developed a system to detect fallen objects on roads from front camera images. Since fallen objects are not specific predefined items, the challenge was exploring methods to detect such unknown objects. Proposed and implemented various detection approaches, from which a primary approach was selected. Implemented and evaluated a novel proposed method, handling everything from data collection and annotation to implementation and evaluation.

Highway Surveillance Camera Image Analysis

Developed a system to detect road anomalies (wrong-way driving, fallen objects, etc.) from object detection results in highway CCTV cameras. Extracted driving paths by analyzing vehicle trajectories using 2D bounding boxes from object detection results, and performed 3D bounding box conversion. Using these techniques, built a system to estimate wrong-way driving and vehicle speed.

DATAGRID Inc.

Part-time

2019.12 - 2020.3 Kyoto

SenseTime Japan, Ltd.

Intern

2019.6 - 2019.9 Kyoto

Hacarus Inc.

Part-time

2019.1 - 2019.4 Kyoto

DENSO Corp.

Intern

2019.2 - 2019.2 Tokyo

Hitachi, Ltd.

Intern

2018.8 - 2018.8 Kanagawa

Show more experiences

Qualification

Computer Graphics Engineer Certification Test (Expert)

Computer Graphic Arts Society (CG-ARTS)

2023.12

2312300081

Image Processing Engineer Certification Test (Expert)

Computer Graphic Arts Society (CG-ARTS)

2023.08

2302400126

TOEIC L&R 815

Educational Testing Service (ETS)

2023.05

JDLA Deep Learning For GENERAL

Japan Deep Learning Association (JDLA)

2018.06

6914272004

Applied Information Technology Engineer Examination

Information-technology Promotion Agency (IPA)

2016.10

AP-2016-10-06502

Fundamental Information Technology Engineer Examination

Information-technology Promotion Agency (IPA)

2014.10

FE-2014-10-09168

Information Technology Passport Examination

Information-technology Promotion Agency (IPA)

2011.10

IP-2011-10-18024

Publications

Dynamic Knowledge Integration in Multi-Agent Systems for Content Inference

* Equal Contribution
Atsushi Yamamoto * , Takumi Iida * , Taito Naruki * , Akihiko Katagiri , Yudai Koike , Ryuta Shimogauchi , Kota Shimomura , Eri Onami , Koki Inoue , Osamu Ito

動的な専門知識連携を意識したマルチエージェントシステム

山本篤 , 成木太音 , 片桐章彦 , 小池湧大 , 飯田啄巳 , 下垣内隆太 , 下村晃太 , 大南英理 , 井上顧基 , 伊藤修
JSAI 2025 2025

マルチモーダルRAGを用いた交通事故リスク推定

千葉 大幹 , 伊藤修 , 飯田啄巳
JSAI 2025 2025

PCT: Perspective Cue Training Framework for Multi-Camera BEV Segmentation

Haruya Ishikawa , Takumi Iida , Yoshinori Konishi , Yoshimitsu Aoki
IROS 2024 (Oral) 2024

LanesPose: 骨格推定によるレーン検出

玄元奏 , 飯田啄巳 , 小西嘉典
MIRU 2023 2023

Detecting Backdoor Attacks on Deep Neural Networks by Saliency Map Clustering

Takumi Iida , Kota Yoshida , Takeshi Fujino
NCSP'20 2020

Bedside Event Real-time Recognition Using Low Resolution Infrared Array Sensor with Recurrent Convolutional Neural Network

Takumi Iida , Kota Yoshida , Takumi Hashimoto , Masayoshi Shirahata , Mitsuru Shiozaki , Takaya Kubota , Masafumi Kimata , Takeshi Fujino
NCSP'18 2018

Patents

落下物検知装置、落下物データ生成方法、および 落下物 学習用データ生成プログラム
饭田啄巳, 松永英树