Computational Photography Patents: Leaders & Filing Trends 2026
- Filing growth has flattened, not accelerated. Filings peaked at 91 in 2023 after climbing from 23 in 2017; the 2022 midpoint of 45 shows the curve was already levelling before the peak, and momentum data shows every major assignee at zero or negative year-over-year growth in the latest year.
- H04N dominates, but G06T is the second center of gravity. 521 of 606 records sit in H04N (pictorial communication) and 297 in G06T (image data processing), meaning most contested claim space overlaps pixel-level signal processing with downstream computational algorithms rather than optics hardware.
- The United States is the primary filing venue by a wide margin. 338 records were filed at the USPTO versus 66 at the EPO, 62 via WIPO/PCT, 55 in China, 17 in Japan and 16 in India — a filing geography worth checking before choosing where to prosecute.
What this patent set covers
This landscape tracks 606 patent families filed against a search built on computational photography, image signal processing and multi-frame fusion, narrowed to claims touching noise reduction, high dynamic range imaging, depth estimation, demosaicing or night mode. The IPC filter (H04N23, G06T5, H04N25) keeps the set focused on pixel-level signal processing and post-capture image transformation rather than lens or sensor hardware alone. Publication typically lags filing by around 18 months, so the most recent filing years in any trend line understate real activity.
The technology composition skews heavily toward H04N and G06T, with meaningful secondary activity in optical systems (G02B), medical imaging (A61B) and AI-based image recognition (G06N, G06V). That spread signals a field where signal-processing claims are increasingly entangled with machine-learning methods for image enhancement, which matters for anyone drafting claims that need to survive both imaging and AI prior art searches.
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 606-family dataset: how filing volume has moved year over year, and how those families distribute across IPC subclasses.
Filing trend, 2017-2026
Filings rose from 23 in 2017 to a peak of 91 in 2023, with the 2022 midpoint at 45 records. The subsequent decline toward 2026 (6 records, partial year) reflects both the publication lag and a genuine flattening that shows up consistently in assignee-level momentum figures.
IPC subclass composition
H04N (521 records) and G06T (297 records) account for the bulk of filings, confirming that pictorial communication and image data processing are the primary battlegrounds. G02B, A61B, G06K, G06V, G06N and G03B each contribute a smaller but non-trivial share, pointing to overlap with optics, medical imaging and AI-driven recognition.
Shares are the percentage of the 606 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Computational Photography and Image Processing with Eureka
This page is one run against one query. Ask Eureka your own question about computational photography and image processing and every answer comes back with the patent numbers behind it.
Try EurekaFoundational and most-cited filings
Image signal processor with adaptive noise reduction and an image signal processing method therefor
The patent describes an image signal processing device that performs correction and recording on input image data while applying noise reduction adapted to imaging conditions and pixel level. A threshold-generating circuit sets the noise threshold using inputting conditions and/or pixel level, and a noise-reducing circuit computes the difference between the subject pixel data and a mean value derived from surrounding pixel data to decide what gets treated as noise.Filed by FUJIFILM CORPORATION, granted 2005-04-19 — an early adaptive noise-reduction architecture referenced throughout later filings in this space.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US8446484B2 | Image processing architecture with pre-scaler | 206 |
| 2 | US20070153335A1 | Image signal processing apparatus, imaging apparatus, image signal processing method and computer program | 197 |
| 3 | US20190197712A1 | Hyperspectral imaging with tool tracking in a light deficient environment | 181 |
| 4 | US20150172539A1 | Distributing processing for imaging processing | 149 |
| 5 | US20190191974A1 | Fluorescence imaging in a light deficient environment | 141 |
| 6 | US20190191975A1 | Fluorescence imaging in a light deficient environment | 131 |
| 7 | US20190191978A1 | Hyperspectral imaging in a light deficient environment | 131 |
| 8 | US20120081567A1 | Techniques for synchronizing audio and video data in an image signal processing system | 131 |
| 9 | US20190191976A1 | Fluorescence imaging in a light deficient environment | 127 |
| 10 | WO2019133739A1 | Fluorescence imaging in a light deficient environment | 126 |
Citation counts favour older filings inside any searched corpus — read them as a measure of influence on later filers, not as a signal of current commercial 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 numbers mean for filing strategy
Three patterns stand out once family counts, IPC composition and momentum are read together.
Growth has already flattened
The climb from 23 filings in 2017 to a peak of 91 in 2023 looks like sustained growth until the midpoint check: 2022 sat at 45, meaning half the peak-year volume had already been reached a year earlier. Combined with a universal drop to zero or negative year-over-year momentum among tracked assignees, this reads as a field consolidating around existing claims rather than expanding into new ones.
Pixel processing and algorithmic fusion overlap heavily
With 521 records in H04N and 297 in G06T out of 606 total, a large share of families likely straddle both subclasses — meaning claims to noise reduction or HDR that only cite one IPC bucket may miss the more crowded intersection. New filings drafted narrowly inside a single subclass risk running into prior art indexed under the other.
US filing dominance shapes freedom-to-operate work
The United States receives more than five times the filings of the EPO and roughly six times China's count. Freedom-to-operate and clearance searches that skip US prosecution history, or treat WIPO/PCT filings (62) as a proxy for global coverage, will miss the bulk of the enforceable claim landscape.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to computational photography and image processing, with the prior art for and against each one.
Who is active, and where activity has stalled
Recent-year momentum tells a different story than cumulative filing counts: the assignees with the largest historical portfolios are, without exception, filing less now than a year ago.
Qualcomm still files, but slower
Qualcomm recorded 2 filings in the latest tracked year, down 80% year over year. It is the only assignee in the momentum data with any filings at all in the most recent period, which makes it the closest thing this dataset has to an active filer.
Major historical filers show zero recent activity
Apple, GoPro, Samsung Electronics, ScanSoft (Nuance-lineage) and Sony all show zero filings in the latest tracked year, with GoPro and Samsung down 100% year over year from the prior period. This does not mean these firms have exited the space — filing pipelines take time to surface — but it does mean their public claim positions are not currently expanding.
A small set of named inventors co-file consistently
The strongest co-assignee pairing in the dataset — Cote Guy and Frederiksen Jeffrey E, appearing together on 9 filings, both also paired separately with Apple on 7 filings each — suggests a stable inventor team operating inside a single corporate program rather than a loose industry collaboration.
| Assignee | Recent year | YoY |
|---|---|---|
| Qualcomm Incorporated | 2 | -80% |
| Apple Inc. | 0 | — |
| GoPro, Inc. | 0 | -100% |
| Samsung Electronics Co., Ltd. | 0 | -100% |
| ScanSoft, Inc. | 0 | — |
| Sony Group Corporation | 0 | — |
| SK Hynix Inc. | 0 | — |
| COTE GUY | 0 | — |
Where to take this analysis
The dataset points to specific next steps depending on whether the goal is clearance, portfolio strategy or identifying open claim territory.
Run a freedom-to-operate check on H04N/G06T overlap
Given that 521 records sit in H04N and 297 in G06T out of 606 total, any new noise-reduction or HDR filing should be checked against both subclasses jointly, not searched separately.
Explore claim overlap in Eureka →Watch Qualcomm's residual filing activity
Qualcomm is the only assignee with any filings in the latest tracked year. Tracking its filings closely may surface where the next wave of claims in this space originates.
Set up assignee tracking in Eureka →Test the under-claimed branches against your own R&D roadmap
Multi-frame depth fusion, adaptive demosaicing and cross-sensor HDR calibration show thinner claim density than core noise-reduction filings — worth validating against internal roadmaps before committing drafting resources.
Draft a scoped search in Eureka →Common questions about this patent landscape
Cumulative filing counts across this 606-family dataset are led by large imaging and mobile-device firms including Apple, Samsung Electronics, Sony and GoPro, alongside chipmakers like Qualcomm. However, recent-year momentum data shows all of these assignees at zero or negative year-over-year filing growth in the latest tracked year, with Qualcomm the only one still filing at all (2 filings, -80% YoY). A historical leadership position does not necessarily indicate current filing activity.
Filing volume grew from 23 records in 2017 to a peak of 91 in 2023, but the 2022 midpoint of 45 shows the growth curve had already flattened before the peak. The drop toward 2026 partly reflects publication lag, since patents typically publish around 18 months after filing, but the consistent negative momentum across every major assignee suggests genuine slowing rather than a reporting artifact alone.
US6882754B2, assigned to Fujifilm and granted in 2005, covers an image signal processor that performs adaptive noise reduction by generating a threshold based on imaging conditions and pixel level, then comparing subject pixel data against a mean derived from surrounding pixels. Its scope is specific to that threshold-generation and mean-comparison mechanism, so it does not block noise-reduction approaches built on different detection logic, such as machine-learning-based denoising models. Anyone filing adaptive noise-reduction claims should still review it as prior art given its age and specificity to threshold-based detection.
The United States is the dominant receiving office in this dataset with 338 records, more than five times the 66 filed at the EPO and well ahead of the 62 filed via WIPO/PCT, 55 in China, 17 in Japan and 16 in India. This concentration means US prosecution history and granted claims carry the most weight for freedom-to-operate work in this field, though EPO and PCT filings still matter for anyone with European or multi-jurisdiction commercial plans.
Relative to the dense core of noise-reduction and HDR claims concentrated in H04N and G06T, sub-areas like multi-frame depth fusion for low-light capture, adaptive demosaicing under mixed illumination, and cross-sensor HDR tone-mapping calibration show thinner representation in this dataset. That does not guarantee an easy grant, but it does suggest less crowded prior art to search against before drafting claims in those specific mechanisms.
Research Computational Photography and Image Processing in depth with Eureka
Go past this page: query the whole computational photography and image processing 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.