Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency
Monocular depth estimation (MDE) is a critical task in computer vision with applications in autonomous driving, robotics, and augmented reality. However, predicting depth from a single image poses significant challenges, especially in dynamic scenes where moving objects introduce scale ambiguity and...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/2/674 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589281250508800 |
|---|---|
| author | Akmalbek Abdusalomov Sabina Umirzakova Makhkamov Bakhtiyor Shukhratovich Azamat Kakhorov Young-Im Cho |
| author_facet | Akmalbek Abdusalomov Sabina Umirzakova Makhkamov Bakhtiyor Shukhratovich Azamat Kakhorov Young-Im Cho |
| author_sort | Akmalbek Abdusalomov |
| collection | DOAJ |
| description | Monocular depth estimation (MDE) is a critical task in computer vision with applications in autonomous driving, robotics, and augmented reality. However, predicting depth from a single image poses significant challenges, especially in dynamic scenes where moving objects introduce scale ambiguity and inaccuracies. In this paper, we propose the Dynamic Iterative Monocular Depth Estimation (DI-MDE) framework, which integrates an iterative refinement process with a novel scale-alignment module to address these issues. Our approach combines elastic depth bins that adjust dynamically based on uncertainty estimates with a scale-alignment mechanism to ensure consistency between static and dynamic regions. Leveraging self-supervised learning, DI-MDE does not require ground truth depth labels, making it scalable and applicable to real-world environments. Experimental results on standard datasets such as SUN RGB-D and KITTI demonstrate that our method achieves state-of-the-art performance, significantly improving depth prediction accuracy in dynamic scenes. This work contributes a robust and efficient solution to the challenges of monocular depth estimation, offering advancements in both depth refinement and scale consistency. |
| format | Article |
| id | doaj-art-a9b38a258961454c809baa6ef9617535 |
| institution | Kabale University |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-a9b38a258961454c809baa6ef96175352025-01-24T13:20:24ZengMDPI AGApplied Sciences2076-34172025-01-0115267410.3390/app15020674Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale ConsistencyAkmalbek Abdusalomov0Sabina Umirzakova1Makhkamov Bakhtiyor Shukhratovich2Azamat Kakhorov3Young-Im Cho4Department of Computer Engineering, Gachon University, Sujeong-gu, Seongnam-si 461-701, Gyeonggi-do, Republic of KoreaDepartment of Computer Engineering, Gachon University, Sujeong-gu, Seongnam-si 461-701, Gyeonggi-do, Republic of KoreaDepartment of Computer Systems, Tashkent University of Information Technologies Named After Muhammad Al-Khwarizmi, Tashkent 100200, UzbekistanDepartment of Artificial Intelligence, Tashkent State University of Economics, Tashkent 100066, UzbekistanDepartment of Computer Engineering, Gachon University, Sujeong-gu, Seongnam-si 461-701, Gyeonggi-do, Republic of KoreaMonocular depth estimation (MDE) is a critical task in computer vision with applications in autonomous driving, robotics, and augmented reality. However, predicting depth from a single image poses significant challenges, especially in dynamic scenes where moving objects introduce scale ambiguity and inaccuracies. In this paper, we propose the Dynamic Iterative Monocular Depth Estimation (DI-MDE) framework, which integrates an iterative refinement process with a novel scale-alignment module to address these issues. Our approach combines elastic depth bins that adjust dynamically based on uncertainty estimates with a scale-alignment mechanism to ensure consistency between static and dynamic regions. Leveraging self-supervised learning, DI-MDE does not require ground truth depth labels, making it scalable and applicable to real-world environments. Experimental results on standard datasets such as SUN RGB-D and KITTI demonstrate that our method achieves state-of-the-art performance, significantly improving depth prediction accuracy in dynamic scenes. This work contributes a robust and efficient solution to the challenges of monocular depth estimation, offering advancements in both depth refinement and scale consistency.https://www.mdpi.com/2076-3417/15/2/674monocular depth estimation (MDE)dynamic scenesscale ambiguityself-supervised learning |
| spellingShingle | Akmalbek Abdusalomov Sabina Umirzakova Makhkamov Bakhtiyor Shukhratovich Azamat Kakhorov Young-Im Cho Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency Applied Sciences monocular depth estimation (MDE) dynamic scenes scale ambiguity self-supervised learning |
| title | Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency |
| title_full | Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency |
| title_fullStr | Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency |
| title_full_unstemmed | Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency |
| title_short | Breaking New Ground in Monocular Depth Estimation with Dynamic Iterative Refinement and Scale Consistency |
| title_sort | breaking new ground in monocular depth estimation with dynamic iterative refinement and scale consistency |
| topic | monocular depth estimation (MDE) dynamic scenes scale ambiguity self-supervised learning |
| url | https://www.mdpi.com/2076-3417/15/2/674 |
| work_keys_str_mv | AT akmalbekabdusalomov breakingnewgroundinmonoculardepthestimationwithdynamiciterativerefinementandscaleconsistency AT sabinaumirzakova breakingnewgroundinmonoculardepthestimationwithdynamiciterativerefinementandscaleconsistency AT makhkamovbakhtiyorshukhratovich breakingnewgroundinmonoculardepthestimationwithdynamiciterativerefinementandscaleconsistency AT azamatkakhorov breakingnewgroundinmonoculardepthestimationwithdynamiciterativerefinementandscaleconsistency AT youngimcho breakingnewgroundinmonoculardepthestimationwithdynamiciterativerefinementandscaleconsistency |