| name | patent-analyst |
| description | 特許分析・知的財産戦略支援スキル。先行技術調査、特許性評価、特許ポートフォリオ分析、
侵害リスク評価、ライセンス戦略策定を支援。特許出願戦略の立案から権利化まで包括的にサポート。
Use when conducting prior art searches, evaluating patentability, analyzing patent portfolios,
assessing infringement risks, or developing IP strategies.
Triggers: "patent search", "prior art", "特許調査", "patentability", "FTO", "特許ポートフォリオ", "IP strategy", "patent landscape".
|
Patent Analyst(特許分析・知的財産戦略)
Overview
This skill provides professional patent analysis and intellectual property (IP) strategy support. Helps organizations conduct prior art searches, evaluate patentability, analyze competitor patent portfolios, assess freedom-to-operate (FTO), and develop comprehensive IP strategies for competitive advantage.
Primary language: Japanese (default), English supported
Patent Databases: USPTO, EPO, JPO, WIPO, Google Patents, Espacenet, PatentScope
Analysis Tools: Patent citation analysis, technology landscape mapping, patent valuation
Output format: Prior art reports, patentability opinions, FTO analysis, IP strategy recommendations, patent landscape reports
When to Use
Use this skill when you need to:
- Conduct novelty searches before patent filing
- Evaluate patentability of inventions
- Perform freedom-to-operate (FTO) analysis before product launch
- Analyze competitor patent portfolios
- Develop patent filing strategies
- Assess patent infringement risks
- Prepare patent prosecution responses (Office Actions)
- Conduct due diligence for M&A or licensing
Trigger phrases (EN): "patent search", "prior art", "patentability", "FTO analysis", "patent portfolio", "IP strategy", "patent landscape", "infringement risk", "Office Action response"
Trigger phrases (JP): "特許調査", "先行技術調査", "特許性評価", "FTO分析", "特許ポートフォリオ", "知的財産戦略", "侵害リスク評価", "拒絶理由通知対応"
Prerequisites
Required Knowledge
- Understanding of patent law basics (35 USC 101, 102, 103, 112 for US; corresponding laws for EP/JP/CN)
- Familiarity with patent classification systems (CPC, IPC)
- Basic Boolean search syntax
Environment Setup
Patent Database Access (Free):
Commercial Tools (Optional, for advanced analysis):
- Derwent Innovation (Clarivate)
- Orbit Intelligence (Questel)
- PatBase (Minesoft)
Python Dependencies (for automation scripts):
pip install requests beautifulsoup4 pandas openpyxl
Workflow
This skill provides four primary workflows:
- Prior Art Search (Novelty Search) - Determine if invention is novel before filing
- Freedom-to-Operate (FTO) Analysis - Identify infringement risks before product launch
- Patent Portfolio Analysis - Assess competitor/acquisition target patent strength
- Patent Prosecution Support - Respond to Office Action rejections
See Core Workflows section for detailed step-by-step procedures.
Quick Start Example
open "https://patents.google.com/?q=neural+network+image+segmentation&type=PATENT"
python3 scripts/patent_search_helper.py \
--keywords "neural network,image segmentation,medical imaging" \
--cpc "G06N3/08,G06V10/82" \
--output prior_art_report.md
Output
This skill generates the following deliverables:
| Output Type | Format | Description |
|---|
| Prior Art Report | Markdown/PDF | Search results, relevance ranking, feature comparison matrix |
| Patentability Opinion | Markdown | 35 USC 102/103 analysis, filing recommendation, risk assessment |
| FTO Analysis Report | Markdown | Claim mapping, infringement risk matrix, design-around options |
| Patent Portfolio Analysis | Markdown/Excel | Technology segmentation, competitor benchmarking, valuation |
| Office Action Response | Markdown | Arguments against rejection, proposed claim amendments |
Sample Output Structure
# Prior Art Search Report
## Executive Summary
- **Invention**: [Title]
- **Search Date**: [Date]
- **Conclusion**: [NOVEL / NOT NOVEL / CONDITIONAL]
## Search Strategy
- **Databases**: Google Patents, Espacenet, USPTO PatFT
- **Keywords**: [List]
- **Classifications**: [CPC codes]
## Key References (Top 10)
| # | Patent/Publication | Relevance | Key Features | Differences |
|---|-------------------|-----------|--------------|-------------|
| 1 | US10,123,456 | HIGH | ... | ... |
## Feature Comparison Matrix
[Detailed comparison table]
## Patentability Assessment
- **Novelty (35 USC 102)**: [Analysis]
- **Non-Obviousness (35 USC 103)**: [Analysis]
## Recommendations
[Filing strategy, claim drafting guidance]
Resources
Bundled References
references/patent_search_methodology.md - Comprehensive search methodology guide
references/claim_analysis_guide.md - Patent claim interpretation and mapping
Bundled Scripts
scripts/patent_search_helper.py - Generate structured prior art search reports
External Resources
Core Framework: Patent Analysis Process
Patent Lifecycle Stages
1. Invention Disclosure
↓
2. Prior Art Search (Novelty Search)
↓
3. Patentability Assessment
↓
4. Patent Application Drafting & Filing
↓
5. Patent Prosecution (Office Actions)
↓
6. Patent Grant
↓
7. Patent Maintenance & Enforcement
↓
8. Monetization (Licensing, Sale, Litigation)
Types of Patent Searches
1. Novelty Search (Patentability Search):
- Purpose: Determine if invention is novel before filing patent application
- Scope: Broad search across all prior art (patents, publications, products)
- Timing: Before filing patent application
- Key Question: "Has this been done before?"
2. Freedom-to-Operate (FTO) Search:
- Purpose: Identify patents that may be infringed by commercializing a product/technology
- Scope: Active patents (in-force) in target markets
- Timing: Before product launch or manufacturing
- Key Question: "Can we practice this technology without infringing third-party patents?"
3. Invalidity Search:
- Purpose: Find prior art to invalidate a competitor's patent (for defense or litigation)
- Scope: Any prior art predating the patent's priority date
- Timing: When threatened with infringement, during litigation, or pre-litigation strategy
- Key Question: "Is there prior art that can invalidate this patent?"
4. State-of-the-Art Search:
- Purpose: Understand technology landscape and trends
- Scope: Broad survey of patents and publications in a technology domain
- Timing: Strategic R&D planning, competitive intelligence
- Key Question: "What is the current state of technology in this field?"
5. Patent Portfolio Analysis:
- Purpose: Analyze competitor or acquisition target's patent portfolio
- Scope: All patents owned by entity (active, expired, pending)
- Timing: Competitive analysis, M&A due diligence, licensing negotiations
- Key Question: "What is the strength and value of this patent portfolio?"
Core Workflows
Workflow 1: Prior Art Search (Novelty Search)
Purpose: Conduct comprehensive search to determine if an invention is novel and non-obvious before filing a patent application.
Step 1: Understand the Invention
Invention Disclosure Meeting:
- Participants: Inventor(s), Patent Attorney/Agent, Patent Analyst
- Duration: 1-2 hours
- Objective: Understand technical details, problem solved, advantages, key features
Key Information to Gather:
## Invention Disclosure Template
### Title
[Descriptive title of invention]
### Technical Field
[E.g., "Machine learning for image recognition" or "Battery electrode materials"]
### Background Problem
- What problem does this invention solve?
- What are the limitations of existing solutions?
- Why are current approaches inadequate?
### Invention Summary (High-Level)
[2-3 sentences describing the core concept]
### Key Features / Novel Elements
1. [Feature 1: e.g., "Use of convolutional neural network with attention mechanism"]
2. [Feature 2: e.g., "Real-time processing on edge device"]
3. [Feature 3: e.g., "Adaptive learning without labeled data"]
### Advantages
- What benefits does this invention provide? (speed, accuracy, cost, efficiency, etc.)
- How does it improve over prior art?
### Embodiments / Variations
- Are there alternative ways to implement this invention?
- Broad vs. narrow interpretations?
### Commercial Application
- What products/services will use this invention?
- Target market and use cases?
### Prior Knowledge
- Are you aware of similar technologies or patents?
- Any relevant publications, products, or conferences?
### Priority Date
- Date of first conception or reduction to practice (for US provisional filing date determination)
Identify Search Concepts:
Break invention into searchable concepts:
- Core Concept: Main technical idea (e.g., "neural network image segmentation")
- Technical Elements: Key components (e.g., "encoder-decoder architecture", "skip connections")
- Application Domain: Use case (e.g., "medical imaging", "autonomous vehicles")
- Functional Aspects: What it does (e.g., "real-time processing", "low latency")
Step 2: Develop Search Strategy
Search Strategy Components:
1. Keywords and Synonyms:
Core Concept: "Image Segmentation"
Synonyms: Image partition, pixel classification, semantic segmentation, instance segmentation
Technical Element: "Convolutional Neural Network"
Synonyms: CNN, deep learning, neural architecture, deep neural network
Application: "Medical Imaging"
Synonyms: Medical image analysis, diagnostic imaging, radiology, CT scan, MRI
2. Patent Classification Codes:
Patent offices use classification systems to categorize inventions:
CPC (Cooperative Patent Classification) - Used by USPTO and EPO:
Example for "Machine Learning for Image Recognition":
- G06N 3/08 (Learning methods - Neural networks)
- G06V 10/82 (Image analysis - Machine learning)
- G06T 7/00 (Image analysis - general)
- A61B 6/00 (Apparatus for radiation diagnosis - if medical imaging)
IPC (International Patent Classification) - Used by WIPO:
Similar to CPC, but broader categories
How to Find Classification Codes:
- Search a few relevant patents and check their classification codes
- Use patent office classification search tools (USPTO CPC, EPO ECLA)
- Browse classification hierarchy to identify related codes
3. Search Query Formulation:
Boolean Operators:
- AND: Both terms must be present
- OR: Either term can be present
- NOT: Exclude term
- NEAR/X: Terms within X words of each other
- Wildcards: * (any characters), ? (single character)
Example Query (Google Patents syntax):
("image segmentation" OR "semantic segmentation")
AND ("neural network" OR CNN OR "deep learning")
AND ("medical imaging" OR "CT scan" OR MRI)
NOT ("text segmentation")
Example Query (USPTO PatFT):
TTL/(image NEAR2 segment$) AND SPEC/(neural network OR deep learning)
AND CPC/(G06N3/08 OR G06V10/82)
Step 3: Conduct Search
Search Databases:
1. Free Databases:
- Google Patents (patents.google.com): Most user-friendly, includes US, EP, JP, WO, CN, KR
- Espacenet (worldwide.espacenet.com): European Patent Office, global coverage
- PatentScope (patentscope.wipo.int): WIPO database, international applications (PCT)
- USPTO PatFT (patft.uspto.gov): US Patent Full-Text database (advanced search)
- J-PlatPat (www.j-platpat.inpit.go.jp): Japan Patent Office (Japanese patents, English abstracts)
2. Commercial Databases (Paid, more powerful):
- Derwent Innovation (Clarivate): Enhanced abstracts, family data, citation analysis
- Orbit Intelligence (Questel): Advanced analytics, patent families, legal status
- PatBase (Minesoft): Comprehensive global coverage, AI-powered search
- LexisNexis TotalPatent: Legal status tracking, alerts
Search Workflow:
1. **Broad Search** (Cast wide net):
- Use core keywords + classification codes
- Review 50-100 patent titles and abstracts
- Identify key patents (most relevant)
2. **Focused Search** (Refine):
- Analyze key patents in detail
- Extract additional keywords, classifications, inventors, assignees
- Perform "forward citation" search (who cited these patents?)
- Perform "backward citation" search (what do these patents cite?)
3. **Deep Dive** (Critical references):
- Full-text review of 10-20 most relevant patents
- Compare claims to your invention
- Identify differences (points of novelty)
4. **Non-Patent Literature (NPL) Search**:
- Google Scholar (scientific publications)
- IEEE Xplore, ACM Digital Library (engineering)
- Product manuals, conference papers, websites
Step 4: Analyze Prior Art
Prior Art Relevance Assessment:
| Reference | Relevance | Key Features Disclosed | Differences from Invention |
|-----------|-----------|------------------------|----------------------------|
| US10,123,456 | High | CNN for image segmentation, encoder-decoder | No attention mechanism, not real-time |
| US9,876,543 | Medium | Real-time image processing | Uses traditional CV (not deep learning) |
| EP3456789 | Medium | Medical image segmentation | Rule-based, no ML |
| Pub: Chen et al. (2018) | High | Attention mechanism in CNN | Not applied to medical imaging, no edge deployment |
**Legend**:
- **High**: Discloses multiple key features of invention
- **Medium**: Discloses some features but missing key elements
- **Low**: Related field but different approach
Feature Comparison Matrix (Invention vs. Prior Art):
| Feature | Our Invention | US10,123,456 | Pub: Chen 2018 | US9,876,543 |
|---------|---------------|--------------|----------------|-------------|
| CNN architecture | ✓ | ✓ | ✓ | ✗ |
| Attention mechanism | ✓ | ✗ | ✓ | ✗ |
| Real-time processing | ✓ | ✗ | ✗ | ✓ |
| Edge device deployment | ✓ | ✗ | ✗ | ✗ |
| Medical imaging application | ✓ | ✓ | ✗ | ✗ |
**Analysis**: No single prior art reference discloses all key features. Combination of US10,123,456 (CNN for medical imaging) + Chen 2018 (attention mechanism) comes closest, but neither teaches edge deployment for real-time processing.
**Conclusion**: Invention appears novel. Potential obviousness rejection based on combination, but teaching away from edge deployment in prior art (computational requirements).
Step 5: Patentability Opinion
Three Criteria for Patentability (US Patent Law 35 USC):
1. Novelty (35 USC 102): Invention must be new
- Question: Is there any single prior art reference that discloses all elements of the invention?
- Assessment: If yes → NOT NOVEL. If no → Proceed to non-obviousness.
2. Non-Obviousness (35 USC 103): Invention must not be obvious to person of ordinary skill in the art
- Question: Would it be obvious to combine prior art references to arrive at the invention?
- Factors (Graham factors):
- Scope and content of prior art
- Differences between prior art and claimed invention
- Level of ordinary skill in the art
- Secondary considerations (commercial success, long-felt need, unexpected results)
- Assessment: Consider whether prior art teaches or suggests combination
3. Utility (35 USC 101): Invention must be useful
- Question: Does the invention have a specific, substantial, and credible utility?
- Assessment: Usually not an issue for most inventions (exception: abstract ideas, laws of nature)
Patentability Opinion Template:
# Patentability Opinion
## Invention Title
[Title]
## Summary of Invention
[2-3 sentences]
## Prior Art Search Summary
- **Databases Searched**: Google Patents, USPTO PatFT, Espacenet, Google Scholar
- **Search Strategy**: Keywords + CPC classifications G06N3/08, G06V10/82
- **Date Range**: All available (focus on last 20 years)
- **References Reviewed**: 85 patents, 15 publications
- **Key References Identified**: 5 highly relevant (listed below)
## Key Prior Art References
1. US10,123,456 (Title: "Deep learning for medical image segmentation") - HIGH relevance
2. Pub: Chen et al., "Attention mechanisms in CNNs" (2018) - HIGH relevance
3. US9,876,543 (Title: "Real-time image processing") - MEDIUM relevance
4. [Additional references...]
## Novelty Analysis (35 USC 102)
**Conclusion**: **NOVEL**
**Rationale**: No single prior art reference discloses all elements of the claimed invention, specifically:
- No reference discloses the combination of:
- CNN with attention mechanism (Element 1)
- Real-time processing on edge device (Element 2)
- Medical imaging application (Element 3)
US10,123,456 discloses Elements 1 and 3 but not Element 2.
Chen 2018 discloses Element 1 but not Elements 2 and 3.
## Non-Obviousness Analysis (35 USC 103)
**Conclusion**: **LIKELY NON-OBVIOUS** (Medium confidence)
**Rationale**:
- **Combination**: It may be argued that US10,123,456 + Chen 2018 renders the invention obvious (both teach attention-based CNNs for image analysis)
- **Teaching Away**: However, prior art teaches that attention mechanisms are computationally expensive and unsuitable for edge devices (see US9,876,543, col. 5, lines 12-18)
- **Unexpected Results**: Our invention achieves real-time performance on edge device despite computational complexity (benchmarks show 10x faster than prior art expectations)
**Risk**: Patent Examiner may issue obviousness rejection based on combination of US10,123,456 and Chen 2018. Response strategy would emphasize:
1. Teaching away from edge deployment in prior art
2. Unexpected performance results
3. Commercial success and long-felt need for real-time medical imaging on portable devices
## Recommendations
**Filing Strategy**: **RECOMMEND FILING** (with considerations)
**Recommended Actions**:
1. **File Patent Application**: Invention appears patentable, proceed with filing
2. **Claim Strategy**:
- Independent claims: Broad claims covering CNN + attention + edge device
- Dependent claims: Specific embodiments (medical imaging, real-time constraints, architecture details)
3. **Potential Prosecution Issues**:
- Anticipate obviousness rejection (combination of US10,123,456 + Chen 2018)
- Prepare evidence of unexpected results (performance benchmarks)
- Consider experimental data to demonstrate non-obviousness
4. **International Filing**: Consider PCT application to preserve foreign filing rights
**Alternative Strategy** (if budget limited):
- Publish invention as defensive publication to prevent others from patenting (creates prior art)
- Maintain as trade secret if embodiment is not easily reverse-engineered (not recommended here since device can be analyzed)
**Estimated Cost** (US filing):
- Patent attorney fees: $8,000 - $15,000
- USPTO filing fees: $1,820 (small entity) / $3,640 (large entity)
- Prosecution (Office Action responses): $3,000 - $8,000
- Total to grant: $15,000 - $30,000
**Timeline to Grant**: 18-36 months (typical for USPTO, depends on art unit backlog)
---
**Prepared By**: [Patent Analyst Name]
**Date**: [Date]
**Review Status**: Preliminary opinion (recommend review by patent attorney before filing)
Workflow 2: Freedom-to-Operate (FTO) Analysis
Purpose: Identify patents that may be infringed by commercializing a product or technology, and assess risk.
FTO vs. Patentability Search
Key Differences:
| Aspect | Patentability Search | FTO Search |
|--------|----------------------|------------|
| **Purpose** | Can we get a patent? | Can we sell this product? |
| **Focus** | Novelty of invention | Risk of infringing others' patents |
| **Prior Art Scope** | All prior art (expired patents, publications, products) | Only ACTIVE patents (in-force) |
| **Geography** | Not critical (novelty is global concept) | CRITICAL (patents are territorial) |
| **Claims Analysis** | Compare invention to prior art | Compare product to patent claims |
| **Timing** | Before filing patent | Before product launch |
Step 1: Define Product/Technology Scope
Product Feature Analysis:
## Product Description: Portable Medical Imaging Device
### Core Features
1. **Hardware**:
- Handheld ultrasound transducer
- Edge computing processor (NVIDIA Jetson)
- Wireless connectivity (Bluetooth 5.0)
- Battery-powered (4-hour runtime)
2. **Software**:
- CNN-based image segmentation algorithm
- Real-time image processing (<100ms latency)
- Cloud sync for image storage
- DICOM format output
3. **Use Case**:
- Point-of-care ultrasound imaging
- Emergency room triage
- Remote/rural healthcare settings
### Technical Specifications
- Image resolution: 512x512 pixels
- Frame rate: 30 FPS
- Weight: 450g
- Regulatory: FDA Class II medical device (510(k) pending)
Step 2: Identify Target Markets
Geographic Scope:
FTO analysis must be conducted separately for each country where product will be manufactured, sold, or used.
Priority Markets (Example):
- United States: Largest medical device market, high litigation risk
- Europe (Germany, UK, France): Significant market, enforceable European Patents (EP)
- Japan: Advanced healthcare system, strong patent protection
- China: Manufacturing location (if applicable), growing market
Note: Patents are territorial (US patent only enforceable in US, EP patent in Europe, etc.)
Step 3: Conduct FTO Search
Search Strategy:
1. Focus on ACTIVE Patents Only:
Active = Patents that are:
- Granted (not just pending applications, though pending should be monitored)
- In-force (maintenance fees paid)
- Not expired (20 years from filing date for utility patents)
Filter:
- Status: Active/In-force
- Priority date: Within last 20 years (typical patent term)
2. Claim-Centric Search (vs. specification search):
FTO analysis focuses on patent CLAIMS (legal definition of invention), not just description.
3. Search Query:
(handheld OR portable OR mobile) AND (ultrasound OR "medical imaging")
AND (edge OR "real-time" OR "image processing")
AND CPC/A61B8/00 (Ultrasound diagnosis)
AND jurisdiction:(US OR EP OR JP OR CN)
AND status:ACTIVE
4. Database Queries:
- Google Patents: Use "Status: Active" filter + jurisdiction filter
- Espacenet: Legal status search (use INPADOC legal status)
- Commercial Tools (Orbit, Derwent): More accurate legal status tracking
Step 4: Claim Mapping Analysis
Claim Mapping Process:
For each potentially relevant patent, map its claims to your product features.
Claim Analysis Template:
## Patent: US10,234,567 (Active, expires 2038)
**Title**: Portable ultrasound device with real-time image processing
**Assignee**: MedTech Corp (competitor)
**Family**: EP, JP, CN equivalents (all active)
### Independent Claim 1 (Simplified)
"A portable medical imaging device comprising:
(a) a handheld ultrasound transducer;
(b) a processor configured to perform real-time image processing;
(c) a wireless communication module; and
(d) a battery power source."
### Claim Mapping to Our Product
| Claim Element | Our Product | Match? | Notes |
|---------------|-------------|--------|-------|
| (a) Handheld ultrasound transducer | ✓ Yes | **YES** | Our product has handheld transducer |
| (b) Processor for real-time image processing | ✓ Yes | **YES** | NVIDIA Jetson performs real-time processing |
| (c) Wireless communication module | ✓ Yes | **YES** | Bluetooth 5.0 module |
| (d) Battery power source | ✓ Yes | **YES** | 4-hour battery |
### Infringement Assessment
**LIKELY INFRINGEMENT** (All claim elements present in our product)
**Risk Level**: **HIGH**
### Potential Defenses / Design-Arounds
1. **Invalidity Challenge**: Search for prior art to invalidate this patent
- Preliminary search found US9,123,456 (published 2 years before priority date) discloses handheld ultrasound with wireless connectivity
- May be able to invalidate Claim 1 based on this prior art
2. **Non-Infringement Arguments**:
- Review dependent claims (narrower scope)
- Analyze claim language (does "processor configured to perform real-time image processing" require specific architecture?)
- Check prosecution history (file wrapper estoppel - did patentee disclaim certain embodiments?)
3. **Design-Around Options**:
- Remove wireless module? (Not practical - key product feature)
- Use tethered design instead of battery? (Defeats portability advantage)
- License from MedTech Corp (recommended approach if patent is valid)
### Licensing Strategy
- **Contact**: MedTech Corp for licensing discussion
- **Estimated Royalty**: 3-5% of product revenue (typical for medical device patents)
- **Negotiation Leverage**: Our patent application may provide cross-licensing opportunity
Step 5: Risk Assessment and Recommendations
FTO Risk Matrix:
| Patent | Assignee | Infringement Risk | Validity Risk | Priority | Recommended Action |
|--------|----------|-------------------|---------------|----------|---------------------|
| US10,234,567 | MedTech Corp | **HIGH** (All elements match) | Medium (potential prior art found) | **P1** | License negotiation + invalidity search |
| US10,456,789 | Siemens | MEDIUM (Most elements, unclear "real-time" threshold) | Low (strong patent) | **P2** | Design-around analysis (avoid "real-time" claim limitations) |
| US10,678,901 | GE Healthcare | LOW (Different ultrasound frequency range) | N/A | **P3** | Monitor (likely non-infringing) |
Risk Level Definitions:
- HIGH: Strong likelihood of infringement, valid patent → Immediate action required
- MEDIUM: Possible infringement, uncertain claim interpretation → Detailed analysis needed
- LOW: Unlikely infringement, significant differences → Monitor, no immediate action
FTO Opinion Summary:
# Freedom-to-Operate Opinion
## Product
Portable Medical Imaging Device (Model: PMI-2025)
## Geographic Scope
United States (primary market)
## Search Summary
- **Patents Reviewed**: 45 active US patents
- **High-Risk Patents Identified**: 1 (US10,234,567)
- **Medium-Risk Patents**: 2 (US10,456,789, US10,789,012)
- **Low-Risk Patents**: 5 (monitoring only)
## Overall Risk Assessment
**MEDIUM-HIGH RISK** (due to US10,234,567)
## Recommended Actions
### Priority 1: US10,234,567 (MedTech Corp)
**Action**: Licensing negotiation
**Rationale**: Likely infringement, licensing preferred over litigation risk
**Timeline**: Initiate contact within 30 days, before product launch
**Estimated Cost**: 3-5% royalty or one-time license fee $500K-$1M
**Alternative**: Conduct invalidity search (budget: $20K-$30K); if successful, may avoid licensing
### Priority 2: US10,456,789 (Siemens)
**Action**: Design-around analysis
**Rationale**: Possible infringement depending on "real-time" claim interpretation
**Timeline**: 60 days (complete before product launch)
**Estimated Cost**: $10K-$15K for detailed claim analysis and design-around engineering
### Priority 3: Monitor Low-Risk Patents
**Action**: Periodic FTO monitoring (quarterly updates)
**Rationale**: Patent landscape changes (new grants, litigation, licensing)
## Product Launch Recommendation
**CONDITIONAL PROCEED** (subject to actions above)
- **Do Not Launch** until Priority 1 (US10,234,567) is resolved (license or invalidity confirmed)
- **Proceed** with design-around for Priority 2 if feasible
- **Risk Tolerance**: If company has $5M+ litigation budget and willingness to challenge patents, may proceed with "design-around + prepared defense" strategy
## Insurance Recommendation
Consider **Patent Infringement Insurance** (Defense cost coverage: $1M-$5M, premiums: $50K-$150K/year)
---
**Prepared By**: [Patent Analyst]
**Date**: [Date]
**Reviewed By**: [Patent Attorney]
**Confidence Level**: Medium-High (based on available claim interpretation, subject to detailed legal analysis)
Workflow 3: Patent Portfolio Analysis
Purpose: Analyze competitor's or acquisition target's patent portfolio to assess strength, value, and strategic positioning.
Step 1: Portfolio Data Collection
Data Points to Collect:
## Portfolio Overview: [Company Name]
### Quantitative Metrics
- **Total Patents**: 1,250 (850 active, 400 expired/abandoned)
- **Pending Applications**: 180
- **Geographic Distribution**:
- US: 450 patents
- EP: 320 patents
- JP: 180 patents
- CN: 200 patents
- Other: 100 patents
- **Filing Trend**: 80-100 new applications per year (last 5 years)
- **Grant Rate**: 75% (industry average: 65%)
### Portfolio Age Distribution
| Age | Count | % |
|-----|-------|---|
| 0-5 years | 320 | 38% |
| 6-10 years | 280 | 33% |
| 11-15 years | 150 | 18% |
| 16-20 years | 100 | 11% |
**Analysis**: Portfolio is relatively young (71% under 10 years old), indicating active R&D and recent innovation focus.
Step 2: Technology Segmentation
Patent Classification by Technology Area:
## Technology Breakdown (by CPC Classification)
| Technology Area | Patents | % of Portfolio | Strategic Importance |
|-----------------|---------|----------------|----------------------|
| Image Processing (G06T) | 380 | 45% | **Core** (company's primary business) |
| Machine Learning (G06N) | 220 | 26% | **Emerging** (growth area) |
| Medical Devices (A61B) | 180 | 21% | **Strategic** (new market entry) |
| User Interfaces (G06F 3) | 70 | 8% | **Supporting** |
**Key Insights**:
- **Concentrated Portfolio**: 45% in image processing (core competency, strong moat)
- **Emerging Bet**: 26% in machine learning (future growth driver)
- **Diversification**: Entering medical device market (21% of portfolio)
Patent Clustering (Topic Modeling):
Use patent analytics tools to cluster patents by technical similarity:
Cluster 1: Real-time image segmentation (150 patents)
Cluster 2: 3D reconstruction from 2D images (90 patents)
Cluster 3: Edge device optimization (80 patents)
Cluster 4: CNN architectures for medical imaging (120 patents)
Cluster 5: Data augmentation techniques (60 patents)
Step 3: Competitor Benchmarking
Competitive Patent Landscape:
| Company | Total Patents | Annual Filings | Technology Focus | Portfolio Strength |
|---------|---------------|----------------|------------------|---------------------|
| **Target Company** | 1,250 | 90/year | Image processing, ML | Strong (Core) |
| Competitor A | 2,800 | 150/year | Broad (Imaging, AI, Robotics) | Very Strong (Diversified) |
| Competitor B | 850 | 60/year | Narrow (Medical imaging only) | Medium (Specialized) |
| Competitor C | 400 | 30/year | Adjacent (Data analytics) | Weak (Few patents) |
**Market Position**: Target Company is #2 in patent count, but #1 in medical imaging sub-segment (180 patents vs. Competitor A's 120 patents in medical).
Patent Citation Analysis:
Forward Citations (Who cites target company's patents?):
- Competitor A: 320 citations (frequent citation indicates fundamental patents)
- Competitor B: 180 citations
- Startups: 150 citations (technology diffusion to ecosystem)
Backward Citations (Who does target company cite?):
- Top cited entities: MIT (80 citations), Stanford (60 citations), IBM (50 citations)
→ Indicates strong academic research foundation and licensing from universities
Step 4: Patent Quality Assessment
Quality Metrics:
1. Claim Breadth:
- Narrow Claims: <5 independent claims, highly specific limitations → Easier to design around
- Broad Claims: Multiple independent claims, broad functional language → Stronger blocking power
Sample Patent Quality Scoring:
| Patent | Independent Claims | Claim Length (words) | Citation Count | Litigation History | Quality Score (1-10) |
|--------|--------------------|-----------------------|----------------|---------------------|----------------------|
| US10,123,456 | 3 | 45 (broad) | 25 (highly cited) | Enforced 2x | **9/10** (High) |
| US10,234,567 | 1 | 120 (narrow) | 5 (low citation) | None | **4/10** (Low) |
| US10,345,678 | 5 | 60 (medium) | 15 | None | **7/10** (Medium-High) |
2. Family Size (Number of countries where patent is filed):
- Large Family (10+ countries): Indicates high strategic value, significant investment
- Small Family (<5 countries): Lower value, defensive filing
3. Maintenance Status:
- Active + Fees Paid: Owner values the patent, likely being practiced
- Abandoned: Owner stopped paying fees → Patent may not be valuable or no longer practiced
Step 5: Valuation and Strategic Assessment
Patent Valuation Approaches:
1. Cost Approach:
Estimated R&D cost to develop patented technology + Patent prosecution costs
Example:
- R&D cost: $500K (engineer salaries, lab equipment, materials)
- Patent costs: $50K (filing + prosecution + maintenance over 10 years)
- Total: $550K per patent
Portfolio (1,250 patents) × $550K = $687M estimated cost
2. Market Approach:
Comparable patent sales or licensing deals in the industry
Example:
- Recent medical imaging patent portfolio (500 patents) sold for $250M
- Per-patent value: $500K
- Target portfolio (1,250 patents, similar quality): $625M estimated value
3. Income Approach (Most common for M&A):
Present value of future cash flows attributable to patents
Example:
- Product revenue attributable to patented technology: $100M/year
- Patent contribution (estimated): 15% of product value = $15M/year
- Discount rate: 10%
- Patent life remaining: 10 years average
- NPV = $15M × (1 - (1.1)^-10) / 0.1 = $92M
→ Portfolio value: $92M (conservative)
Strategic Value Assessment:
## Strategic Value (Beyond Financial)
### Offensive Value (Assertion Potential)
- **High**: 180 medical imaging patents cover emerging market (competitors likely infringe)
- **Medium**: 220 ML patents (broad field, harder to enforce)
- **Low**: 70 UI patents (design-arounds available)
### Defensive Value (Freedom to Operate)
- **Essential Patents**: 50 patents likely essential for anyone practicing real-time medical imaging
- **Blocking Power**: Strong portfolio prevents competitors from entering medical imaging without license
### Licensing Revenue Potential
- **Current**: $12M/year in licensing revenue (8 active licensees)
- **Potential**: $30M/year if aggressively licensed (estimated 20 additional licensees in market)
### Cross-Licensing Leverage
- Strong portfolio enables favorable cross-licensing deals with competitors (avoids cash payments)
### Acquisition Strategic Fit
- **If Acquiring**: Portfolio complements acquirer's existing patents (25% overlap, 75% additive)
- **If Target of Acquisition**: Portfolio is key asset (justifies 30-40% of acquisition premium)
Workflow 4: Patent Prosecution Support (Office Action Response)
Purpose: Respond to patent examiner's Office Action rejections to overcome objections and obtain patent grant.
Common Rejection Types
1. 35 USC 102: Novelty Rejection:
- Meaning: Examiner found prior art that discloses all elements of claimed invention
- Response Strategy: Argue differences, amend claims to distinguish from prior art
2. 35 USC 103: Obviousness Rejection:
- Meaning: Examiner argues combination of prior art renders invention obvious
- Response Strategy: Argue non-obviousness (teaching away, unexpected results, commercial success)
3. 35 USC 112: Enablement/Written Description Rejection:
- Meaning: Specification does not adequately describe or enable the invention
- Response Strategy: Argue specification discloses invention, or amend claims to match disclosure
4. 35 USC 101: Subject Matter Eligibility Rejection (Software/Business Methods):
- Meaning: Claimed invention is directed to abstract idea, law of nature, or natural phenomenon
- Response Strategy: Argue practical application, technological improvement
Office Action Response Template
# Response to Office Action
**Application Number**: 17/123,456
**Filing Date**: March 15, 2023
**First Named Inventor**: John Smith
**Title**: Real-Time Image Segmentation on Edge Device Using Attention-Based CNN
**Examiner**: Jane Doe
**Art Unit**: 2624
---
## Overview of Examiner's Rejections
The Office Action dated October 10, 2024 rejected Claims 1-15 under the following grounds:
1. **Claims 1, 2, 5-10**: Rejected under 35 USC 103 as obvious over US10,123,456 (Johnson) in view of US10,234,567 (Chen)
2. **Claims 3-4, 11-15**: Rejected under 35 USC 102(a)(1) as anticipated by US10,345,678 (Lee)
Applicant respectfully traverses these rejections and requests reconsideration.
---
## Response to 35 USC 103 Rejection (Claims 1, 2, 5-10)
### Examiner's Position (Summary)
The Examiner argues that it would have been obvious to combine:
- Johnson (US10,123,456): Discloses CNN-based image segmentation for medical imaging
- Chen (US10,234,567): Discloses attention mechanisms in CNNs
The Examiner states: "It would have been obvious to incorporate Chen's attention mechanism into Johnson's CNN segmentation system to improve accuracy." (Office Action, p. 3)
### Applicant's Response
**Respectfully, the Examiner's conclusion is not supported for the following reasons:**
#### Reason 1: Teaching Away from Combination
Johnson expressly teaches AWAY from using computationally expensive mechanisms like attention on edge devices:
> "Attention mechanisms, while effective, are unsuitable for resource-constrained edge devices due to their computational complexity (>10 GFLOPS), which exceeds typical edge processor capabilities (1-5 GFLOPS)." (Johnson, col. 8, lines 45-50)
This clear teaching away from the claimed combination contradicts the Examiner's obviousness rationale. One of ordinary skill in the art would be **discouraged** from combining Johnson and Chen based on Johnson's explicit guidance.
#### Reason 2: Unpredictable Results
The claimed invention achieves unexpected results not suggested by the prior art:
**Performance Comparison** (Attached as Exhibit A - Declaration of Dr. Smith):
| System | Inference Time (ms) | Accuracy (%) | Device |
|--------|---------------------|--------------|--------|
| Johnson (alone) | 120 | 88% | Edge device |
| Chen (alone) | 250 | 92% | Server GPU |
| **Claimed Invention** | **85** | **94%** | **Edge device** |
The claimed invention achieves:
- **Faster inference** than Johnson alone (85ms vs. 120ms) while incorporating attention (contradicting computational complexity concern)
- **Higher accuracy** than Johnson (94% vs. 88%)
- **Edge device deployment** unlike Chen (which requires server GPU)
These results are unexpected and would not have been predictable to one of ordinary skill in the art at the time of invention.
#### Reason 3: Secondary Considerations
**Commercial Success**: Since product launch (June 2024), Applicant has sold 5,000 units with >$10M revenue, capturing 15% market share in portable medical imaging segment.
**Long-Felt Need**: Prior to Applicant's invention, no commercially available edge-based real-time medical imaging device with CNN segmentation existed. The market had sought such a device for over 5 years (see market reports attached as Exhibit B).
**Industry Recognition**: Applicant's invention received "Most Innovative Medical Device" award from Medical Technology Association (2024) and has been cited in 12 subsequent patent applications by competitors.
These secondary considerations support non-obviousness of the claimed invention.
### Conclusion on 103 Rejection
For the foregoing reasons, Applicant respectfully submits that Claims 1, 2, 5-10 are not obvious over Johnson in view of Chen. The prior art teaches away from the claimed combination, the invention produces unexpected results, and secondary considerations strongly support non-obviousness.
**Respectfully request reconsideration and withdrawal of rejection.**
---
## Response to 35 USC 102 Rejection (Claims 3-4, 11-15)
### Examiner's Position
The Examiner asserts that Lee (US10,345,678) anticipates Claims 3-4, 11-15 by disclosing all elements of the claimed invention.
### Applicant's Response
**Respectfully, Lee does NOT anticipate the claimed invention** because Lee fails to disclose at least the following claim element:
**Disputed Claim Element (Claim 3)**:
> "wherein the attention mechanism is configured to dynamically weight feature maps based on spatial and channel-wise importance..."
**Lee's Disclosure** (cited by Examiner at col. 5, lines 10-20):
> "The system applies a fixed weighting matrix to feature maps to emphasize edge regions..."
**Analysis**: Lee's "fixed weighting matrix" is fundamentally different from the claimed "dynamically weight feature maps based on spatial and channel-wise importance":
| Aspect | Claimed Invention | Lee |
|--------|-------------------|-----|
| Weighting | **Dynamic** (changes per input) | **Fixed** (pre-defined matrix) |
| Basis | **Spatial AND channel-wise** | **Spatial only** (edge regions) |
| Adaptation | **Input-dependent** | **Static** |
A person of ordinary skill in the art would understand these as distinct approaches. Lee's fixed weighting matrix does not read on the claimed dynamic attention mechanism.
### Proposed Claim Amendment (If Necessary)
If the Examiner maintains the rejection, Applicant is willing to amend Claim 3 to further clarify:
**Amended Claim 3** (changes underlined):
> "wherein the attention mechanism is configured to **adaptively and dynamically** weight feature maps **in real-time** based on spatial and channel-wise importance **computed from input-specific features**..."
However, Applicant submits that this amendment is not necessary, as the original claim language is sufficiently distinct from Lee.
### Conclusion on 102 Rejection
For the foregoing reasons, Lee does not anticipate Claims 3-4, 11-15. **Respectfully request withdrawal of rejection.**
---
## Conclusion and Request
For all the reasons set forth above, Applicant respectfully submits that Claims 1-15 are patentable and requests reconsideration and withdrawal of all rejections.
Should the Examiner have any questions or wish to discuss this application, Applicant's attorney is available for an Examiner interview.
**Respectfully submitted,**
[Patent Attorney Name]
Registration No. [Number]
[Law Firm]
[Contact Information]
**Date**: November 15, 2024
---
**Attachments**:
- Exhibit A: Declaration of Dr. John Smith (Inventor) with performance benchmarks
- Exhibit B: Market reports evidencing long-felt need
- Exhibit C: Highlighted portions of Johnson and Chen references
Best Practices
1. Conduct Thorough Prior Art Searches
Don't skip the prior art search:
- Filing a patent without prior art search wastes money if invention is not novel
- Prior art search informs claim drafting strategy (avoid known art)
- Budget: $2,000-$5,000 for professional prior art search (worth the investment)
2. File Early, File Strategically
Priority Date is Critical:
- US: First-to-file system (not first-to-invent) - file ASAP after invention
- Consider provisional application ($1,800 filing fee) to establish priority date, then convert to non-provisional within 12 months
Strategic Filing:
- PCT (Patent Cooperation Treaty) application: Preserves foreign filing rights for 30 months while deferring costs
- Direct filing in key markets: US, EU, JP, CN (if budget allows)
3. Focus on Claim Quality
Claims Define the Invention (and enforce ability):
- Broad independent claims (capture core invention, harder for competitors to design around)
- Narrow dependent claims (fallback if broad claims are rejected)
- Functional language (what invention does) + structural language (how it's built)
4. Monitor Competitor Patents
Competitive Intelligence:
- Set up alerts for competitor patent publications (Google Patents, Espacenet alerts)
- Conduct annual FTO reviews (patent landscape changes)
- Track competitor litigation and licensing activities
5. Build a Strategic Portfolio
Portfolio Strategy:
- Core Patents: Protect key products and technologies (high value, defend at all costs)
- Blocking Patents: Prevent competitors from entering your space (may not practice yourself)
- Peripheral Patents: Improvements and variations (incremental value)
- Defensive Publications: Publish inventions you don't want to patent (creates prior art, prevents competitors from patenting)
Common Pitfalls
❌ Filing Too Late
Invention publicly disclosed before filing patent → Public disclosure becomes prior art (bars patentability in most countries except US 1-year grace period).
Solution: File provisional application BEFORE any public disclosure (conference, publication, product launch).
❌ Ignoring FTO Analysis
Launching product without FTO analysis → Risk of patent infringement lawsuit after significant investment.
Solution: Conduct FTO analysis 6-12 months before product launch, while design changes are still feasible.
❌ Abandoning Valuable Patents
Letting patents expire due to non-payment of maintenance fees → Lose exclusive rights.
Solution: Implement patent docket management system, review portfolio annually, prune only truly low-value patents.
❌ Over-Drafting Claims
Filing patent with only narrow claims → Competitors easily design around.
Solution: Include multiple claim scopes (broad independent + narrow dependent), cover embodiments and variations.
❌ Disclosing Trade Secrets in Patents
Filing patent on technology that should be trade secret → 20-year protection vs. indefinite trade secret protection (if reverse-engineering is difficult).
Solution: Reserve patents for inventions that will be publicly known (products), use trade secrets for processes and algorithms not visible to public.
このスキルの目的は、組織の知的財産戦略を支援し、特許調査、特許性評価、FTO分析、特許ポートフォリオ分析を通じて、競争優位性を確保し、特許侵害リスクを最小化することです。適切な特許戦略により、イノベーションを保護し、ビジネス価値を最大化してください。