Optical Flow and Motion Estimation Patents: Leaders & Trends 2026
- Filing peaked in 2025 at 159 records after a flat-to-declining stretch through the 2022 midpoint (59), suggesting the field surged late rather than growing steadily.
- China accounts for 416 of 795 families more than the United States (162), Europe (60) and WIPO (60) combined, concentrating receiving-office activity heavily in one jurisdiction.
- Only 10 co-assignee pairs across 795 families co-filing is rare here — most assignees are building their optical flow claims independently rather than through joint ventures.
Filing growth compares 2021 (55 records) with 2024 (72) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 795 records in scope (CR5), not by the ranked leaders only.
What this landscape covers
Optical flow and motion estimation patents in this dataset span 795 patent families filed between 2015 and mid-2026, spanning classic displacement-field methods like Lucas-Kanade and Farneback alongside neural-network-based dense flow estimation. The search combines terminology filters (“optical flow”, “motion estimation”, “occlusion handling”, “temporal consistency”) with IPC codes across G06T7 (image data processing), G06V10 (image/video recognition) and H04N19 (video coding), capturing both the algorithmic core and its video-compression and surgical-imaging applications.
Because publication lags filing by roughly 18 months, the 2026 count of 40 understates actual filing activity for that year; the 2025 peak of 159 is the most reliable recent-year signal. Family-level counting is used throughout rather than raw document counts, which reduces the distortion from continuation filings and multi-jurisdiction refiling of the same invention.
Let an AI agent run this analysis on your own technology
Pick a task. Every answer cites the patents behind it.
Filing trends and technology composition
Two views of the same 795-family dataset: how filing volume has moved year over year, and which IPC subclasses carry the claim weight.
A late-cycle surge, not steady growth
Filings held roughly flat from 2017 (48) through the 2022 midpoint (59) before climbing sharply to a 2025 peak of 159. That shape points to a recent wave of interest — likely tied to neural-network-based dense flow methods — rather than a technology that has been scaling steadily for a decade.
Image processing dominates, video coding trails
G06T (image data processing & generation) covers 627 of 795 records and G06V (image/video recognition) covers 316, confirming that most activity treats optical flow as a core image-processing problem. H04N (video/TV pictorial communication) at 152 and the smaller A61B (diagnosis & surgery, 29) and G01B (length measurement, 19) counts show the application layer — video compression, surgical navigation, dimensional measurement — is comparatively thin.
Shares are the percentage of the 795 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Optical Flow and Motion Estimation with Eureka
This page is one run against one query. Ask Eureka your own question about optical flow and motion estimation and every answer comes back with the patent numbers behind it.
Try EurekaThe most-cited records and a representative filing
US20220383521A1 — Dense optical flow calculation system and method based on FPGA
Disclosed are a dense optical flow calculation system and method based on an FPGA (Field Programmable Gate Array). The system comprises a software system deployed on a host and a dense optical flow calculation module deployed on the FPGA. Pixel information of two continuous frames of pictures is obtained from a host end in the system, and optical flow is obtained by calculation by means of the steps such as smoothing processing, polynomial expansion, intermediate variable calculation, optical flow calculation. An image pyramid and iterative optical flow calculation can be achieved by repeatedly calling a calculation core module in the FPGA; a final calculation result is returned to the host.Filed by Zhejiang University, published 2022-12-01 — an example of hardware-accelerated classical (Farneback-style) optical flow rather than a neural approach.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US6385245B1 | Motion estimation and motion-compensated interpolation | 355 |
| 2 | US20140369584A1 | Method And Apparatus For Determining Tumor Shift During Surgery Using A Stereo-Optical Three-Dimensional Surf… | 278 |
| 3 | US6278736B1 | Motion estimation | 243 |
| 4 | US20190138889A1 | Multi-frame video interpolation using optical flow | 170 |
| 5 | US20200265590A1 | Methods, systems, and computer readable media for estimation of optical flow, depth, and egomotion using neur… | 145 |
| 6 | US5600731A | Method for temporally adaptive filtering of frames of a noisy image sequence using motion estimation | 118 |
| 7 | US6295377B1 | Combined spline and block based motion estimation for coding a sequence of video images | 96 |
| 8 | US20160191795A1 | Method and system for presenting panoramic surround view in vehicle | 84 |
| 9 | US20190355128A1 | Segmenting generic foreground objects in images and videos | 82 |
| 10 | US9076201B1 | Volumetric deformable registration method for thoracic 4-D computed tomography images and method of determini… | 73 |
Citation counts favour older records simply because they have had more time to accumulate citations inside this corpus — read them as a signal of historical influence, not current relevance.
Each row carries its publication number; clicking a row searches Eureka by that number.
Put your own technology through the same analysis
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →What the data means for filing strategy
Three patterns stand out once volume, geography and citation weight are read together.
China is the primary filing venue
With 416 of 795 families routed through China against 162 for the US and 60 each for Europe and WIPO, most recent activity — including the FPGA-based dense optical flow filing from Zhejiang University — is being staked out domestically first.
The highest-cited prior art predates the neural-network wave
US6385245B1, cited 355 times, and US6278736B1, cited 243 times, both address classical motion-compensated interpolation and estimation — foundational blocking art that any new displacement-field claim still has to clear.
Filers build claim positions alone
Only 10 co-assignee pairs appear across the whole corpus, and the strongest is a single company-plus-inventor pairing rather than a cross-company alliance, indicating this space is largely staked out through solo filing rather than joint development.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to optical flow and motion estimation, with the prior art for and against each one.
| Assignee | Co-assignee | Shared families |
|---|---|---|
| Koninklijke Philips N.V. (Royal Philips) | DE HAAN GERARD | 3 |
| Bracco Imaging S.p.A. | WERNER VOMWEG TONI | 2 |
| Bracco Imaging S.p.A. | MATTIUZZI MARCO | 2 |
| Bracco Imaging S.p.A. | GORI ILARIA | 2 |
| Koninklijke Philips N.V. (Royal Philips) | Philips Norden AB | 1 |
| Koninklijke Philips N.V. (Royal Philips) | WITTERBROOD RIMMERT B | 1 |
| Koninklijke Philips N.V. (Royal Philips) | WITTEBROOD RIMMERT B | 1 |
| Koninklijke Philips N.V. (Royal Philips) | PHILIPS AB | 1 |
The strongest co-assignee link (Koninklijke Philips N.V. (Royal Philips) with inventor DE HAAN GERARD, 3 shared filings) is an inventor relationship inside one company, not an inter-company alliance — there is little evidence of joint-venture patenting in this field.
Who is filing, and where momentum has stalled
Recent-year filing counts by assignee show a field where even the more active historical filers have slowed sharply.
Qualcomm is the only assignee still filing at pace
Among the tracked assignees, Qualcomm Incorporated (Qualcomm) is the sole name with any filing in the latest year (1); Koninklijke Philips N.V. (Royal Philips) (Philips), Magic Leap, Inc. (Magic Leap), Meta Platforms, Inc. (Meta Platforms) and Bracco Imaging S.p.A. (Bracco Imaging) all show zero.
Samsung's filing has dropped to zero
Samsung Electronics Co., Ltd. (Samsung Electronics) posted a -100% year-over-year change with zero filings in the latest year, a pattern consistent with the broader post-2025-peak pullback visible across the assignee set.
Older motion-compensation art still anchors the field
The most-cited record in the corpus, on motion estimation and motion-compensated interpolation, sits well ahead of any recent filing in citation count, underlining how foundational the pre-neural-network era remains for freedom-to-operate analysis.
| Assignee | Recent year | YoY |
|---|---|---|
| Qualcomm Incorporated | 1 | — |
| Koninklijke Philips N.V. (Royal Philips) | 0 | — |
| Magic Leap, Inc. | 0 | — |
| Meta Platforms, Inc. | 0 | — |
| Bracco Imaging S.p.A. | 0 | — |
| Samsung Electronics Co., Ltd. | 0 | -100% |
| Johns Hopkins University | 0 | — |
| Canon Inc. | 0 | — |
Where to take this analysis
The dataset points to two practical next steps for teams making filing or freedom-to-operate decisions in this space.
Model the post-2025 pullback
Filing volume dropped from a 2025 peak of 159; confirming whether 2026's partial count of 40 reflects genuine slowdown or publication lag is the first thing worth checking before reading too much into the recent trend.
Explore filing trends in EurekaMap claims against the under-claimed branches
Event-based sensing, occlusion-aware surgical displacement fields and hardware-accelerated pipelines show thinner filing density than the core G06T image-processing claims — each is worth a targeted prior-art pull before drafting.
Run a white space search in EurekaCommon questions about optical flow and motion estimation patents
The most-cited record in this corpus is US6385245B1, on motion estimation and motion-compensated interpolation, with 355 citations, followed by US6278736B1 on motion estimation with 243. These are older, foundational patents rather than recent filings, which is typical of citation data — older records simply have had more time to accumulate citations. For current market position, filing volume and recent-year momentum by assignee are a better guide than citation count alone.
Filing volume was roughly flat from 2017 (48) through the 2022 midpoint (59), then rose sharply to a peak of 159 in 2025. The 2026 figure of 40 is partial and understated because publication typically lags filing by about 18 months, so it should not be read as a decline yet. The overall shape looks like a late-cycle surge tied to neural-network-based dense flow methods rather than steady multi-year growth.
China leads by a wide margin with 416 of 795 families, more than the United States (162), the European Patent Office (60) and WIPO PCT filings (60) combined. Japan (27) and Israel (18) are smaller but present. Anyone assessing freedom to operate in this field should weight Chinese-language prior art and CNIPA filings heavily, not just US and EPO records.
US20220383521A1, filed by Zhejiang University and published 2022-12-01, describes a dense optical flow calculation system implemented on an FPGA, using smoothing, polynomial expansion, and iterative pyramid-based calculation — essentially a hardware-accelerated version of Farneback-style dense optical flow. It matters because it sits in the thinly-populated intersection of classical dense flow algorithms and dedicated hardware acceleration, an area with less filing density than pure software or neural-network claims. Anyone building FPGA- or ASIC-accelerated optical flow pipelines should review its claim scope before finalizing an architecture.
IPC composition shows heavy concentration in G06T (627 of 795 records) and G06V (316), with thinner coverage in application-specific areas like A61B surgical imaging (29) and G01B dimensional measurement (19). Sub-areas such as event-based optical flow sensing, occlusion-aware displacement fields for surgical navigation, and hardware-accelerated dense flow pipelines show comparatively less filing density than the core image-processing claims. These under-claimed branches are worth a dedicated prior-art search before drafting new applications.
Research Optical Flow and Motion Estimation in depth with Eureka
Go past this page: query the whole optical flow and motion estimation corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
Disclaimer. This page is generated from Patsnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. Patsnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company’s registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.