https://taisei-ando.github.io/publications/cubedvo/Mon, 01 Jun 2026 00:00:00 +0000https://taisei-ando.github.io/publications/cubedvo/<div class="cubedvo-page-marker"></div> <p><strong>Links:</strong> / / </p> <h2 id="abstract">Abstract</h2> <p>This paper proposes CubeDVO, a learning-based visual odometry system specifically designed for a monocular 360-degree camera. While recent methods apply spherical convolutions to equirectangular images, their core operations still rely on the distorted 2D domain, which inherently degrades tracking accuracy. To address this limitation, our method introduces a unified cubemap-spherical representation that decouples the pose estimation pipeline from the equirectangular image space. We extract features on cubemap planes to reduce projection-induced distortion while enabling standard 2D convolutions, and perform geometric optimization on the unit sphere for consistent pose estimation. Our architecture incorporates a geometry-aware flow estimation network and a differentiable spherical bundle adjustment module. CubeDVO is evaluated against state-of-the-art spherical and conventional perspective visual odometry methods. Experimental results demonstrate that our method achieves robust performance under aggressive motion while maintaining high accuracy and practical computational efficiency.</p> <h2 id="keywords">Keywords</h2> <p>visual odometry, omnidirectional vision, spherical geometry, deep learning</p>https://taisei-ando.github.io/experience/Tue, 24 Oct 2023 00:00:00 +0000https://taisei-ando.github.io/experience/https://taisei-ando.github.io/bib/cubedvo/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/bib/cubedvo/https://taisei-ando.github.io/bib/equirectangular-2025/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/bib/equirectangular-2025/https://taisei-ando.github.io/bib/iccas-2024-ando/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/bib/iccas-2024-ando/https://taisei-ando.github.io/bib/self-tio/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/bib/self-tio/https://taisei-ando.github.io/bib/tc-ltio/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/bib/tc-ltio/https://taisei-ando.github.io/cv/Mon, 01 Jan 0001 00:00:00 +0000https://taisei-ando.github.io/cv/<div class="cv-page-marker"></div> <h2 id="basics">Basics</h2> <ul class="cv-basics-list"> <li><strong>Name:</strong> Taisei Ando</li> <li><strong>Affiliation:</strong> Ph.D. Student in Robotics, The University of Tokyo</li> <li><strong>Email:</strong> ando [at] robot.t.u-tokyo.ac.jp</li> </ul> <h2 id="education">Education</h2> <div class="cv-timeline"> <div class="cv-timeline-item"> <p class="cv-timeline-title"><span class="cv-timeline-main">Ph.D. Student, The University of Tokyo</span><span class="cv-timeline-detail">Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences</span></p> <p class="cv-timeline-date">Apr. 2026 - Present</p> <p>Advisor: Prof. Atsushi Yamashita</p> </div> <div class="cv-timeline-item"> <p class="cv-timeline-title"><span class="cv-timeline-main">M.S., The University of Tokyo</span><span class="cv-timeline-detail">Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences</span></p> <p class="cv-timeline-date">Apr. 2024 - Mar. 2026</p> <p>Advisor: Prof. Atsushi Yamashita</p> </div> <div class="cv-timeline-item"> <p class="cv-timeline-title"><span class="cv-timeline-main">B.E., The University of Tokyo</span><span class="cv-timeline-detail">Department of Precision Engineering, Faculty of Engineering</span></p> <p class="cv-timeline-date">Apr. 2020 - Mar. 2024</p> <p>Advisor: Associate Prof. Qi An</p> </div> </div> <h2 id="work">Work</h2> <div class="cv-timeline"> <div class="cv-timeline-item"> <p class="cv-timeline-title"><span class="cv-timeline-main">Research Assistant, National Institute of Advanced Industrial Science and Technology (AIST)</span><span class="cv-timeline-detail">Smart Mobility Research Team, Intelligent System Research Institute</span></p> <p class="cv-timeline-date">Jan. 2026 - Present</p> <p>Advisor: Dr. Kenji Koide</p> </div> <div class="cv-timeline-item"> <p class="cv-timeline-title"><span class="cv-timeline-main">Data Analysis Intern, D-Stats</span></p> <p class="cv-timeline-date">Apr. 2023 - May 2026</p> <p>Responsibilities: coding, code review, and mentoring</p> </div> </div> <h2 id="honors">Honors</h2> <ul> <li> <p><strong>Partial Repayment Exemption for Outstanding Achievement, Type 1 Graduate Scholarship</strong><br> Jul. 2026, Japan Student Services Organization (JASSO)</p> </li> <li> <p><strong>Graduate School Research Encouragement Award</strong><br> Feb. 2026, Society of Automotive Engineers of Japan (JSAE)</p> </li> <li> <p><strong>Excellent Presentation Award</strong><br> Dec. 2025, The 26th SICE System Integration Division Annual Conference (SI2025), The Society of Instrument and Control Engineers (SICE)</p> </li> <li> <p><strong>Best Paper Award</strong><br> Oct. 2024, The 24th International Conference on Control, Automation and Systems (ICCAS 2024)</p> </li> <li> <p><strong>Fellowship</strong><br> Oct. 2024 - Present, International Graduate Program of Innovation for Intelligent World, The University of Tokyo</p> </li> <li> <p><strong>Scholarship</strong><br> Apr. 2022 - Mar. 2026, Satomi Scholarship Foundation</p> </li> </ul> <h2 id="skills">Skills</h2> <div class="cv-skills-grid"> <div> <p><strong>Programming Languages</strong></p> <ul> <li>Python</li> <li>C++</li> <li>C</li> <li>CUDA</li> </ul> </div> <div> <p><strong>Libraries and Frameworks</strong></p> <ul> <li>PyTorch</li> <li>OpenCV</li> <li>ROS</li> <li>Git/GitHub</li> <li>Docker</li> </ul> </div> </div> <h2 id="miscellaneous">Miscellaneous</h2> <ul> <li><strong>Reviewer Service</strong> <ul> <li>IEEE Robotics and Automation Letters (IEEE RA-L)</li> </ul> </li> <li><strong>Application Development</strong> <ul> <li> (App Store)</li> <li> (App Store)</li> </ul> </li> </ul>