PRESS-NEWS.org - Press Release Distribution
PRESS RELEASES DISTRIBUTION

MoBluRF: A framework for creating sharp 4D reconstructions from blurry videos

Researchers develop new framework that can create sharp neural radiance fields from blurry monocular videos, captured from everyday handheld devices

2025-09-19
(Press-News.org)

Neural Radiance Fields (NeRF) is a fascinating technique that creates three-dimensional (3D) representations of a scene from a set of two-dimensional (2D) images, captured from different angles. It works by training a deep neural network to predict the color and density at any point in 3D space. To do this, it casts imaginary light rays from the camera through each pixel in all input images, sampling points along those rays with their 3D coordinates and viewing direction. Using this information, NeRF reconstructs the scene in 3D and can render it from entirely new perspectives, a process known as novel view synthesis (NVS).

Beyond still images, a video can also be used, with each frame of the video treated as a static image. However, existing methods are highly sensitive to the quality of the videos. Videos captured with a single camera, such as those from a phone or drone, inevitably suffer from motion blur caused by fast object motion or camera shake. This makes it difficult to create sharp, dynamic NVS. This is because most existing deblurring-based NVS methods are designed for static multi-view images, which fail to account for global camera and local object motion. In addition, blurry videos often lead to inaccurate camera pose estimations and loss of geometric precision.

To address these issues, a research team jointly led by Assistant Professor Jihyong Oh from the Graduate School of Advanced Imaging Science (GSIAM) at Chung-Ang University (CAU) in Korea, and Professor Munchurl Kim from Korea Advanced Institute of Science and Technology (KAIST), Korea, along with Mr. Minh-Quan Viet Bui, Mr. Jongmin Park, developed MoBluRF, a two-stage motion deblurring method for NeRFs. “Our framework is capable of reconstructing sharp 4D scenes and enabling NVS from blurry monocular videos using motion decomposition, while avoiding mask supervision, significantly advancing the NeRF field,” explains Dr. Oh. Their study was made available online on May 28, 2025, and was published in Volume 47, Issue 09 of the IEEE Transactions on Pattern Analysis and Machine Intelligence on September 01, 2025.

MoBluRF consists of two main stages: Base Ray Initialization (BRI) and Motion Decomposition-based Deblurring (MDD). Existing deblurring-based NVS methods attempt to predict hidden sharp light rays in blurry images, called latent sharp rays, by transforming a ray called the base ray. However, directly using input rays in blurry images as base rays can lead to inaccurate prediction. BRI addresses this issue by roughly reconstructing dynamic 3D scenes from blurry videos and refining the initialization of “base rays” from imprecise camera rays.

Next, these base rays are used in the MDD stage to accurately predict latent sharp rays through an Incremental Latent Sharp-rays Prediction (ILSP) method. ILSP incrementally decomposes motion blur into global camera motion and local object motion components, greatly improving the deblurring accuracy. MoBluRF also introduces two novel loss functions, one that separates static and dynamic regions without motion masks, and another that improves geometric accuracy of dynamic objects, two areas where previous methods struggled.

Owing to this innovative design, MoBluRF outperforms state-of-the-art methods with significant margins in various datasets, both quantitatively and qualitatively. It is also robust against varying degrees of blur.

“By enabling deblurring and 3D reconstruction from casual handheld  captures, our framework enables smartphones and other consumer devices to produce sharper and more immersive content,” remarks Dr. Oh. “It could also help create crisp 3D models of shaky footages from museums, improve scene understanding and safety for robots and drones, and reduce the need for specialized capture setups in virtual and augmented reality.”

MoBluRF marks a new direction for NeRFs, enabling high quality 3D reconstructions from ordinary blurry videos recorded with everyday devices.

 

***

 

Reference

DOI: https://doi.org/10.1109/TPAMI.2025.3574644

 

About Chung-Ang University

Chung-Ang University is a leading private research university in Seoul, South Korea, dedicated to shaping global leaders for an evolving world. Founded in 1916 and achieving university status in 1953, it combines academic tradition with a strong commitment to innovation. Fully accredited by the Ministry of Education, CAU excels in fields such as pharmacy, medicine, engineering, and applied sciences, driving impactful discoveries and technological progress. Its research-intensive environment fosters collaboration and excellence, producing scholars and professionals who lead in their disciplines. Committed to global engagement, CAU continues to expand its influence as a hub for scientific advancement and future-driven education.

Website: https://neweng.cau.ac.kr/index.do

About Assistant Professor Jihyong Oh

Dr. Jihyong Oh is an Assistant Professor in the Department of Imaging Science, Graduate School of Advanced Imaging Science, Multimedia & Film (GSAIM) at Chung-Ang University (CAU; Seoul, South Korea), and has been leading the Creative Vision and Multimedia Lab (https://cmlab.cau.ac.kr/) since September 2023. Prior to this, he was a postdoctoral researcher at VICLAB at KAIST (Daejeon, South Korea). He received his B.E., M.E., and Ph.D. degrees in Electrical Engineering from KAIST in 2017, 2019, and 2023, respectively. His research primarily focuses on low-level vision, image/video restoration, 3D vision, and generative AI. He has published papers at CVPR, ICCV, ECCV, AAAI, TPAMI, TCSVT, Neurocomputing, and Remote Sensing, and serves as a reviewer for these conferences and journals, as well as for SIGGRAPH and IEEE TPAMI, TIP, TGRS, TMM and Access. He received Outstanding Reviewer Awards for ICCV 2021 and CVPR 2024.

Website: https://scholarworks.bwise.kr/cau/researcher-profile?ep=1528

END



ELSE PRESS RELEASES FROM THIS DATE:

Community management effectively protects millions of hectares of Amazonian forest

2025-09-19
New research reveals “unprecedented” conservation results of community-based management of protected areas in the Amazon - as many face a future in which they may become increasingly degraded due to low enforcement of regulations, growing external encroachment and competition for resources. The study describes a powerful new mechanism for increasing the extent of effective area-based protection by piggybacking on community management of natural resources. Tropical protected areas are typically understaffed, underfunded and underequipped and it remains unclear how existing ones ...

No single solution for gastroparesis — patients need personalized care

2025-09-19
Bethesda, MD (Sept. 17, 2025) — The American Gastroenterological Association (AGA) has released a new clinical guideline with 12 conditional recommendations for diagnosing and managing gastroparesis, a serious and often debilitating disease. Rather than pointing to a single treatment path, the evidence highlights a range of options — underscoring the complexity of gastroparesis and the importance of thoughtful conversations between doctors and patients to tailor an individualized care plan.  Gastroparesis affects thousands of people, often causing nausea, postprandial fullness, vomiting, discomfort, and major disruptions to quality of life. The ...

Stakeholders meet to discuss national peatland impact plans for Finland, Germany, Netherlands

2025-09-19
Between 2nd  and 10th September around 40 stakeholders involved in protecting and restoring peatlands and wetlands in Germany, the Netherlands and Finland joined a series of online workshops to develop impact plans that aim to influence policies and markets in each country, to scale up the restoration of degraded peatlands.  The meetings were organised by the EU-funded research project WET HORIZONS, the UNEP’s Global Peatlands Initiative, Wetlands International, and Eurosite - the European Land Conservation Network, who are coordinating the European Peatlands Initiative.  “The ...

Physically cold, mentally strained

2025-09-19
Performance in endurance sports, such as long-distance running, cycling, and cross-country skiing, requires sustained activity over prolonged periods of time, and is influenced by multiple factors, such as temperature and mental state. In chilly temperatures around 10°C (50°F), endurance performance has been shown to improve in certain circumstances. In contrast, extreme cold can slow down muscle and nerve function and cause a decline in performance. Such reactions suggest the involvement of the Sympathetic Adrenal-Medullary (SAM) system, a stress response pathway. Further performance decline can be attributed to mental fatigue caused by psychological pressure, which leads to decreased ...

Consistent policy, not “patchwork” regulations, recommended for the coexistence of crops

2025-09-19
Australian farmers face inconsistent guidelines when it comes to crop regulations across genetically modified (GM), organic and other crop frameworks, according to new research from the University of Adelaide. “Even though different sectors in Australian cropping regulate coexistence of both genetically modified and organic crops, they do so in different ways,” says lead researcher Michail Ivanov, whose review was published in Griffith Law Review. “For example, different standards or codes of conduct recommend different physical barriers or buffer zones between paddocks to ...

LEDs shed light on efficient tomato cultivation

2025-09-19
Researchers including those from the University of Tokyo have successfully grown large tomatoes and cherry tomatoes, both rich in nutrients, in tightly controlled environments where the light source was energy-efficient LEDs. Such methods were often limited by the types or sizes of plants that could thrive in such conditions. This feasibility study demonstrates the researchers’ method is suitable for urban environments, potentially even in space, and can offer food security in the face of climate change or extreme weather conditions. Pizza, pasta, soup, salad, the tomato ...

2025 Ig Physics Nobel Prize for perfect pasta sauce

2025-09-19
The Ig Nobel Prize honors research that first makes people laugh, then makes them think. Its 35th award ceremony possibly also makes people hungry: ISTA physicist Fabrizio Olmeda and colleagues researched the secret of a perfect cacio e pepe pasta sauce. They received the popular award for their findings on Thursday evening in Boston, USA. Cacio e pepe is one of Italy's most popular pasta dishes, but even (Italian) scientists often fail to prepare the perfect creamy sauce. Fabrizio Olmeda, a physicist at the Institute of Science and Technology Austria (ISTA), also struggled ...

Bright squeezed light in the kilohertz frequency band

2025-09-19
Bright squeezed light, exhibiting sub-shot-noise quantum noise combined with significant optical power, is essential for enhanced sensitivity in quantum metrology and precision measurement. In the deep application within the field, the squeezing must be extended to kHz–MHz bandwidth with milliwatt optical power. However, it remains a longstanding challenge to achieve this goal with conventional technologies, since they suffer from low-frequency technical noise and vacuum noise coupling, which caps ...

Water flowed on ancient asteroid

2025-09-19
A team of researchers, including those at the University of Tokyo, discovered that liquid water once flowed on the asteroid that spawned near-Earth asteroid Ryugu more than a billion years after it first formed. The finding, based on tiny rock fragments returned by the Hayabusa2 spacecraft of the Japan Aerospace Exploration Agency (JAXA), overturns long-held assumptions that water activity on asteroids only occurred in the earliest moments of solar system history. This could impact models which include the formation of the Earth. We have a relatively good ...

AI model offers accurate and explainable insights to support autism assessment

2025-09-19
Scientists have developed and tested a deep-learning model that could support clinicians by providing accurate results and clear, explainable insights – including a model-estimated probability score for autism. The model, outlined in a study published in eClinicalMedicine (a journal from The Lancet), was used to analyse resting-state fMRI data – a non-invasive method that indirectly reflects brain activity via blood-oxygenation changes. In doing so, the model achieved up to 98% cross-validated accuracy for Autism Spectrum ...

LAST 30 PRESS RELEASES:

New SCAI initiative targets faster shock diagnosis and care with ‘door to lactate clearance’

MoBluRF: A framework for creating sharp 4D reconstructions from blurry videos

Community management effectively protects millions of hectares of Amazonian forest

No single solution for gastroparesis — patients need personalized care

Stakeholders meet to discuss national peatland impact plans for Finland, Germany, Netherlands

Physically cold, mentally strained

Consistent policy, not “patchwork” regulations, recommended for the coexistence of crops

LEDs shed light on efficient tomato cultivation

2025 Ig Physics Nobel Prize for perfect pasta sauce

Bright squeezed light in the kilohertz frequency band

Water flowed on ancient asteroid

AI model offers accurate and explainable insights to support autism assessment

Process for dealing with sexual misconduct by doctors requires major reform

Severe pregnancy sickness raises risk of mental health conditions by over 50%

Early humans may have walked from Türkiye to mainland Europe, new groundbreaking research suggests

New study shows biochar’s electrical properties can influence rice field methane emissions

Guangdong faces largest chikungunya outbreak on record

Tirzepatide improves blood sugar control in children aged 10-17 years with type 2 diabetes inadequately controlled on existing therapies (SURPASS-PEDS trial)

An old drug, in a low dose, shown to be safe and effective in preventing progression of type 1 diabetes in children and young people (MELD-ATG trial)

Study reports potential effects of verapamil in slowing progression of type 1 diabetes

Fresh hope for type 1 diabetes as daily pill that slows onset confirms promise at 2-year follow-up

New estimates predict over 4 million missing people who would be alive in 2025 if not for inadequate type 1 diabetes care

So what should we call this – a grue jay?

Chicago Quantum Exchange-led coalition advances to final round in NSF Engine competition

Study identifies candidates for therapeutic targets in pediatric germ cell tumors

Media alert: The global burden of CVD

Study illuminates contributing factors to blood vessel leakage

What nations around the world can learn from Ukraine

Mixing tree species does not always make forests more drought-resilient

Public confidence in U.S. health agencies slides, fueled by declines among Democrats

[Press-News.org] MoBluRF: A framework for creating sharp 4D reconstructions from blurry videos
Researchers develop new framework that can create sharp neural radiance fields from blurry monocular videos, captured from everyday handheld devices