Computational Imaging Patents: Who Leads, Where the Gaps Are 2026
- Concentrated at the top. The top 5 assignees hold 75.0% of all 24 records in scope, and the top 10 hold 95.8% — this is a field with very little room left among the leaders.
- Filing has cooled from its peak. Activity peaked at 5 records in 2020; the most recent complete comparison, 2021 to 2024, shows a -67% change, though 2025-26 figures are still filling in under publication lag.
- Image processing dominates the claims. G06T (image data processing) appears in 66.7% of records and H04N (pictorial communication) in 58.3%, while AI-model classes and recognition classes each sit at 4.2-12.5%.
Filing growth compares 2021 (3 records) with 2024 (1) — 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 24 records in scope (CR5), not by the ranked leaders only.
What this landscape covers
This review covers 24 published records matching computational photography and multi-frame image fusion, filtered against claims on motion artefacts, noise reduction, high dynamic range, multi-camera depth, processing latency and on-device power. The scope spans 2015 through the 2026 data cut-off, and captures the shift from single-frame HDR blending toward motion-corrected, depth-aware fusion pipelines built for mobile and embedded capture.
Because publication lags filing by roughly 18 months, the trend line for the last one to two years is necessarily incomplete — read the most recent years as a floor, not a ceiling, on actual filing activity.
Filing trend and technology composition
Two views of the same 24-record set: how filing activity has moved year over year, and which technology classes carry the claims.
Filing trend, 2017-2026
Filings rose to a peak of 5 records in 2020, then eased; the 2021-to-2024 comparison (3 down to 1, a -67% change) is the most recent span long enough to read with confidence, since 2025 and 2026 are still being backfilled by publication lag.
IPC subclass composition
G06T (image data processing and generation) and H04N (pictorial communication) between them cover the great majority of records at 66.7% and 58.3% respectively, since most filings combine an image-processing pipeline claim with a video/transmission claim. Smaller shares in G06N, G01N, G06K and G06V mark AI-model, material-analysis and recognition angles that sit alongside the core fusion claims rather than replacing them; note these shares are measured against all 24 records and sum to more than 100% because a single record can carry several classes.
Shares are the percentage of the 24 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Computational Imaging and Multi-Frame Fusion with Eureka
This page is one run against one query. Ask Eureka your own question about computational imaging and multi-frame fusion and every answer comes back with the patent numbers behind it.
Try EurekaRepresentative and most-cited filings
Image alignment for computational photography (US20220345605A1)
Image frames for computational photography may be corrected, such as through rolling shutter correction, prior to fusion of the frames to reduce wobble and jitter artefacts in a video sequence of HDR-enhanced images. First and second motion data are determined for the capture times of first and second frames, rolling shutter correction is applied based on both sets of motion data, and the corrected frames are then aligned and fused into a single output frame with higher dynamic range.Filed by Qualcomm; published 2022-10-27.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20170127046A1 | Depth Masks for Image Segmentation for Depth-based Computational Photography | 56 |
| 2 | CN111292264A | 一种基于深度学习的图像高动态范围重建方法 | 47 |
| 3 | US8160392B1 | Automatic suggestion of image blending techniques | 21 |
| 4 | CN106162131A | 一种实时图像处理方法 | 14 |
| 5 | US20220414834A1 | Computational photography features with depth | 13 |
| 6 | CN110060219A | 一种基于低秩近似的真实图降噪方法 | 11 |
| 7 | CN113724146A | 基于即插即用先验的单像素成像方法 | 7 |
| 8 | US20220345605A1 | Image alignment for computational photography | 6 |
| 9 | US10554956B2 | Depth masks for image segmentation for depth-based computational photography | 4 |
| 10 | CN119064368A | 一种基于图像识别的建筑施工污染度检测方法 | 2 |
Citation counts are drawn from within this searched corpus and favour older filings that have had more time to accumulate citations — treat them as a signal of influence rather than a ranking of current importance.
Patent titles are shown in the language they were filed in, not translated, so that each record stays verifiable against the original filing — a translated title will not match in Eureka or in any national register. 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 numbers say
Three read-outs from the filing and citation data that matter for deciding where to file next.
A short leader group holds most of the field
The top 5 assignees account for 75.0% of all 24 records in scope, and the top 10 push that to 95.8%. With only 11 companies in the ranking overall, this is a field where a small number of players occupy nearly the entire claim space.
Filing has slowed from its 2020 peak
After peaking at 5 records in 2020, filings fell across the last fully-comparable window, 2021 to 2024, by 67%. Years after 2024 are not yet reliable for a trend read because of publication lag.
Image-processing and video-communication classes dominate
G06T and H04N together cover the bulk of filings, meaning most protection sits on the image-pipeline and transmission side rather than on the AI-model or recognition side, where filing density remains thin at 4.2-12.5% of records.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to computational imaging and multi-frame fusion, with the prior art for and against each one.
Who holds the claim space
The ranking below is the entire assignee list the dataset returns — 11 companies across 24 records — not a filtered top slice.
One assignee sets the pace
The leading assignee holds 6 of the 24 records in scope, the largest single share in a field where the top 5 combined already cover 75.0% of all filings.
A tight cluster below the leader
Fifth place holds 2 records, showing a step down from the leader rather than a long gradual taper — most of the meaningful activity sits inside a handful of names.
A thin tail of single-filing entrants
By tenth place, holdings drop to a single record, and the top 10 combined still only reach 95.8% of the 24-record total — there is very little filing activity outside the ranked leaders.
| Assignee | Recent year | YoY |
|---|---|---|
| Qualcomm Incorporated | 0 | — |
| Google LLC | 0 | — |
| Ai Tao | 0 | — |
| Dell Products L.P. | 0 | — |
| Beijing Institute of Technology | 0 | — |
| Shenzhen Fuse Technology Co., Ltd. | 0 | -100% |
| Wuhan University | 0 | — |
| Microsoft Technology Licensing, LLC | 0 | — |
Where to take this analysis
The dataset points to a concentrated leader group and a thin set of under-claimed branches — the next step is deciding where a new filing would actually clear.
Check freedom-to-operate against the leaders
With 75.0% of records held by five assignees, any new filing on core HDR fusion or motion-corrected alignment should be checked against those portfolios first.
Explore assignee portfolios in Eureka →Look at the under-claimed branches
AI-model-driven fusion and recognition-integrated capture each sit at 4.2-12.5% of records, well below the core image-processing classes — worth a closer claim-scope read before assuming the space is occupied.
Run a white-space search in Eureka →Common questions on this landscape
The dataset ranks 11 companies across 24 published records, and the field is concentrated: the top 5 assignees together hold 75.0% of all records, and the top 10 hold 95.8%. The single leading assignee holds 6 of the 24 records outright. This means a small group of filers occupy most of the claim space, and any freedom-to-operate check should start with that leader group rather than the long tail.
Filings peaked at 5 records in 2020 and then declined; comparing 2021 (3 records) to 2024 (1 record) shows a -67% change over that span. Filing counts for 2025 and 2026 are still incomplete because publication typically lags actual filing by around 18 months, so those most recent years should not yet be read as confirming a continued decline. The safest read is that activity cooled after 2020, with the trend beyond 2024 still unresolved.
Most records combine an image-data-processing claim (G06T, 66.7% of the 24 records) with a pictorial-communication or video claim (H04N, 58.3%), reflecting pipelines that process and then transmit or display the fused frame. Smaller shares cover AI-model computation, material analysis, and image/video recognition, each at 4.2% to 12.5% of records. These shares add to more than 100% because individual filings often carry more than one IPC class.
This Qualcomm filing claims correcting image frames — specifically rolling shutter correction based on motion data captured at the time each frame was taken — before fusing the frames into a single HDR output. It targets a specific sequence: per-frame motion data, motion-based correction, then alignment and fusion. Anyone building a motion-corrected HDR fusion pipeline for mobile or video capture should read this claim scope closely, since it sits at the centre of the field's densest citation cluster.
The core image-processing and video-communication classes are heavily claimed, at 66.7% and 58.3% of the 24 records respectively, so new filings there face dense prior art. Branches with much thinner coverage include AI-model-driven fusion, material or sensor-analysis integration, and recognition-integrated capture, each sitting at only 4.2% to 12.5% of records. These lower-density branches are worth a closer look before assuming the space is already occupied, though thin filing density is not proof the underlying technology is easy to design around.
Research Computational Imaging and Multi-Frame Fusion in depth with Eureka
Go past this page: query the whole computational imaging and multi-frame fusion 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.