Capsule Endoscope AI Patents: Leaders & Where Filings Slowed 2026
- Filing peaked in 2022 at 7 families, then flattened — the dataset does not show a technology still accelerating, it shows one that has settled into its claim positions.
- Image processing dominates the IPC mix, with A61B and G06T covering the great majority of records while G10L speech analysis and H04N video pictorial claims sit at just 2 records each.
- The United States and India lead filing offices, at 21 and 11 records respectively, ahead of WIPO PCT, China and South Korea — a split worth checking before choosing where to file first.
What this landscape covers
Capsule endoscope AI and machine learning patents cover the intersection of ingestible imaging hardware and the software that reads what it captures: lesion and bleeding detection, image classification, and the diagnostic-support pipelines built on top. The corpus searched here spans 44 patent families published between 2015 and the 2026 cut-off, filtered to records that combine capsule endoscope hardware language with machine-learning or classification claim language under IPC classes A61B1/04, G06T7 and G16H50.
Publication lags filing by roughly 18 months, so the most recent filing year in this dataset is understated by construction — treat 2025 and 2026 counts as a floor, not a ceiling.
Filing trend and technology mix
Two views of the same 44 families: how filing volume moved year over year, and which IPC subclasses carry the claim language.
A flat-to-declining filing curve
Filings rose from 2 in 2017 to a peak of 7 in 2022, then did not sustain that pace through to the 2026 cut-off. Read against an 18-month publication lag, this looks less like a technology in decline and more like one whose visible filing rate has plateaued — the open question is whether 2024–2026 filings still pending publication will lift the recent years once they clear the pipeline.
Image processing carries the claim weight
A61B (diagnosis & surgery, 39 records) and G06T (image data processing, 30 records) between them touch nearly every family in the set. G06N (AI models, 13), G06V (image/video recognition, 12) and G16H (healthcare informatics, 11) form a secondary tier built on that imaging base, while G10L (speech/audio, 2) and H04N (pictorial communication, 2) are thin enough to be genuinely open ground rather than crowded niches.
Shares are the percentage of the 44 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Capsule Endoscope AI and Machine Learning with Eureka
This page is one run against one query. Ask Eureka your own question about capsule endoscope ai and machine learning and every answer comes back with the patent numbers behind it.
Try EurekaThe most-cited records in this corpus
Methods, programs, apparatus for obtaining health information from sensors in an ingestible capsule
The claim covers deriving a spectral analysis from a motion sensor's time-series data as an ingestible capsule passes through the GI tract, and using that spectral analysis to detect peristalsis at a location within the tract during that passage.Filed by ATMO BIOSCIENCES LIMITED, published 2025-08-28 as WO2025175357A1.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20190125173A1 | Method and apparatus for tracking position of capsule endoscope | 28 |
| 2 | US20150065850A1 | Accurate and efficient polyp detection in wireless capsule endoscopy images | 28 |
| 3 | WO2021054477A2 | Disease diagnostic support method using endoscopic image of digestive system, diagnostic support system, diag… | 25 |
| 4 | US20180365826A1 | Capsule endoscope for determining lesion area and receiving device | 24 |
| 5 | US20180308235A1 | SYSTEM and METHOAD FOR PREPROCESSING CAPSULE ENDOSCOPIC IMAGE | 22 |
| 6 | CN107292347A | 一种胶囊内窥镜图像识别方法 | 21 |
| 7 | US20200065614A1 | Endoscopic image observation system, endosopic image observation device, and endoscopic image observation met… | 11 |
| 8 | KR101875004B1 | Automated bleeding detection method and computer program in wireless capsule endoscopy videos | 11 |
| 9 | US20220022736A1 | Systems and methods for collecting and screening of pancreatic secretions | 10 |
| 10 | CN109934276A | 基于迁移学习的胶囊内窥镜图像分类系统及方法 | 7 |
Citation counts inside a searched corpus favour older records by construction — read this table as a map of technical influence to date, not of current commercial 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 mean for a filing decision
Three findings that shape where a new filing would sit relative to existing claim density.
Growth has not continued past the 2022 peak
The midpoint year of the dataset is also its high point on record so far. Combined with the publication lag, this means the technology's real 2024–2026 trajectory is not yet visible — but the visible curve gives no basis for claiming acceleration.
Imaging and diagnosis claims dominate the corpus
Nearly every family touches image data processing or core diagnosis/surgery claim language. Secondary AI-specific classes (G06N, G06V) sit at roughly a third of that volume, suggesting many filings bundle machine-learning claims inside broader imaging or device claims rather than filing AI methods standalone.
Filing is concentrated in two offices, not spread evenly
The United States and India together account for the large majority of receiving-office activity, with WIPO PCT, China and South Korea trailing well behind. A filer targeting Europe or Japan directly would be filing into comparatively thin prior art by office, though PCT routes may still surface family members there.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to capsule endoscope ai and machine learning, with the prior art for and against each one.
Who holds the claim space
Recent-year momentum across the tracked assignees shows no filer adding new families in the latest year — a signal to check publication lag before reading it as retreat.
No assignee shows fresh filings in the latest tracked year
Every assignee in the momentum data — spanning device makers, research institutes and university tech-transfer arms — shows zero families in the most recent year, including one recording a -100% year-on-year change. Given the roughly 18-month publication lag, this is consistent with a normal reporting gap rather than an actual stop in R&D.
A mix of device manufacturers and research institutions
The tracked assignees include capsule-endoscope hardware makers, a national electronics research institute, and university industry-cooperation groups — indicating that both commercial device firms and academic labs are actively claiming in this space, not just one type of filer.
The earliest influential claims are on tracking and detection
The two most-cited records address capsule position tracking and polyp detection in capsule images — foundational problems that later filings on lesion detection and diagnostic support appear to build on.
| Assignee | Recent year | YoY |
|---|---|---|
| Ankon Technologies (Wuhan) Co., Ltd. | 0 | — |
| Electronics and Telecommunications Research Institute (ETRI) | 0 | — |
| INDIRA GANDHI DELHI TECH UNIV FOR WOMEN (IGDTUW) | 0 | — |
| Given Imaging Ltd. | 0 | -100% |
| Olympus Corporation | 0 | — |
| Ajou University Industry-Academic Cooperation Foundation | 0 | — |
| SUNGSHIN WOMENS UNIV IND ACADEMIC COOP FOUND | 0 | — |
| SRM INST OF SCI & TECH | 0 | — |
Using this landscape in a filing or freedom-to-operate decision
The dataset points to specific next checks rather than a single conclusion.
Verify the 2024–2026 gap before acting on it
Because publication lags filing by about 18 months, the apparent plateau after 2022 needs re-checking against a fresh pull once more of the recent filing years have published.
Explore filing trends in EurekaMap claim language against the thin IPC classes
G10L and H04N sit at only 2 records each in this corpus — worth a closer read to confirm whether that reflects genuine white space or simply classification choices by other filers.
Run a claim-level search in EurekaCheck the ingestible sensor route independently
WO2025175357A1's motion-sensor and spectral-analysis approach sits outside the core imaging classes that dominate this corpus, and may warrant its own prior-art search rather than being read as part of the imaging cluster.
Trace citing families in EurekaCommon questions on capsule endoscope AI patents
This landscape identifies 44 patent families published between 2015 and the mid-2026 data cut-off, matched against capsule endoscope hardware terms combined with machine-learning and classification claim language under IPC classes A61B1/04, G06T7 and G16H50. That is a modest but well-defined corpus rather than a sprawling one, which makes individual filings easier to review directly. Because publication lags filing by roughly 18 months, the true count for 2024 through 2026 filings will rise as those applications clear examination and publish.
The visible trend rose from 2 families in 2017 to a peak of 7 in 2022, and has not exceeded that peak in the years published since. That does not necessarily mean interest has cooled — the 18-month publication lag means recent years are structurally undercounted — but the data gives no basis for claiming continued acceleration. Anyone using this trend for a go/no-go decision should treat 2025 and 2026 figures as a floor and revisit the count after a fresh pull.
A61B (diagnosis and surgery) and G06T (image data processing) are the two dominant IPC subclasses, appearing in 39 and 30 of the 44 records respectively — meaning most filings touch core imaging or diagnostic claim language. AI-specific classes such as G06N and G06V appear less often, at 13 and 12 records, suggesting many patents fold machine-learning claims into broader imaging or device claims rather than filing them as standalone AI methods. Speech/audio analysis (G10L) and video pictorial communication (H04N) are the thinnest classes at just 2 records each.
Based on IPC composition, the thinnest classes in this corpus are G10L (speech and audio analysis) and H04N (pictorial communication for video), each at only 2 records against 39 for the leading A61B class. That points toward audio-based motility sensing, multi-sensor fusion beyond optical imaging, and real-time video compression for capsule feeds as areas with comparatively less claim density. A proper freedom-to-operate check should still search each branch directly rather than relying on subclass counts alone.
The United States leads receiving offices with 21 records, followed by India at 11, then WIPO's PCT route at 6, China at 4 and South Korea at 2. That concentration in the US and India suggests those are the primary markets filers are prioritizing for protection, while direct filings into China, South Korea, and other jurisdictions remain comparatively sparse in this dataset. PCT filings may still extend family coverage into additional countries not captured in this receiving-office breakdown.
Research Capsule Endoscope AI and Machine Learning in depth with Eureka
Go past this page: query the whole capsule endoscope ai and machine learning 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.