Efficient measurement and reconstruction techniques for various occluded scenes

2023 Fiscal Year Final Research Report

• Project/Area Number: 22K21283
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: Keio University
• Principal Investigator: Isogawa Mariko 慶應義塾大学, 理工学部(矢上), 准教授 (60963238)
• Project Period (FY): 2022-08-31 – 2024-03-31
• Keywords: Non-Line-of-Sight / 被遮蔽領域 / センシング / シーン復元

Outline of Final Research Achievements

In this project, we tackled the measurement and scene reconstruction methods for occluded Non-Line-of-Sight (NLOS) scenes, which cannot be directly observed by cameras or sensors. Specifically, we aimed to develop technology that allows the measurement of areas which previously required densely placed sensors, using only a limited number of sensors. To this end, we tackled this project in two directions: "A: Scene measurement and reconstruction methods based on non-aligned and sparse measurement points" and "B: NLOS imaging applicable to non-planar relay walls." The results were presented at three domestic conferences (two presented, one scheduled for presentation) and submitted to the international conference and to a journal paper (under review).

Free Research Field

Computer vision

Academic Significance and Societal Importance of the Research Achievements

被遮蔽領域の計測や復元手法には，曲がり角越し状態推定等の自動運転車向けの危険予測への応用や，瓦礫の中から被遮蔽領域の状況を計測しそれを救護活動に活用すること，監視カメラの死角となってしまう領域においてもシーンの状況を把握可能とするセキュリティ用途で用いることなど，大きな社会的ニーズが期待できる．本研究では特に，従来よりも計測点が少ない場合においても良好に被遮蔽シーンを復元可能な円形計測に基づく手法や，非平面な中継壁にも活用可能な被遮蔽シーン復元手法の構築を目指した高速な画像修復技術の適用可能性を検討した．これらの成果により，被遮蔽シーン復元技術の実用可能性を向上させられたと考えている．