GITNUXSOFTWARE ADVICE
Technology Digital MediaTop 10 Best Face Software of 2026
Compare the top 10 Face Software tools for face recognition accuracy and features. Explore picks and get the right fit fast.
How we ranked these tools
Core product claims cross-referenced against official documentation, changelogs, and independent technical reviews.
Analyzed video reviews and hundreds of written evaluations to capture real-world user experiences with each tool.
AI persona simulations modeled how different user types would experience each tool across common use cases and workflows.
Final rankings reviewed and approved by our editorial team with authority to override AI-generated scores based on domain expertise.
Score: Features 40% · Ease 30% · Value 30%
Gitnux may earn a commission through links on this page — this does not influence rankings. Editorial policy
Editor’s top 3 picks
Three quick recommendations before you dive into the full comparison below — each one leads on a different dimension.
Google Cloud Vision AI
Face detection with landmarks plus emotion and safety signals in a single Vision request
Built for teams building automated face and document vision pipelines with managed APIs.
Amazon Rekognition
Liveness detection that flags likely spoofing during face verification
Built for teams building AWS-native face recognition workflows with managed APIs.
Microsoft Azure Face
Face identification and verification using persisted face lists and similarity-based matching
Built for applications needing face detection and attribute analysis with Azure identity features.
Related reading
Comparison Table
This comparison table side-by-side evaluates Face Software options that provide face recognition, detection, and related vision features across major cloud platforms and specialized APIs. Readers can scan key capability differences, such as supported face tasks, indexing and search workflows, latency and deployment models, integration approach, and typical use cases for identity verification, fraud detection, and media analytics.
| # | Tool | Category | Overall | Features | Ease of Use | Value |
|---|---|---|---|---|---|---|
| 1 | Google Cloud Vision AI Vision AI provides face detection and face-related analysis through Google Cloud APIs. | API-first | 9.2/10 | 9.4/10 | 9.3/10 | 8.9/10 |
| 2 | Amazon Rekognition Rekognition offers face detection and face search capabilities through AWS APIs. | API-first | 8.9/10 | 8.8/10 | 8.9/10 | 9.2/10 |
| 3 | Microsoft Azure Face Azure Face exposes face detection and face recognition features via REST APIs on Microsoft Azure. | API-first | 8.6/10 | 9.0/10 | 8.4/10 | 8.3/10 |
| 4 | Clarifai Clarifai provides face detection and face recognition models through hosted machine learning APIs. | API-platform | 8.3/10 | 8.4/10 | 8.4/10 | 8.2/10 |
| 5 | AWS Amazon Face Search AWS face search features for identifying faces are available through Rekognition services documented for production use. | enterprise API | 8.0/10 | 8.3/10 | 7.9/10 | 7.8/10 |
| 6 | IBM Watson Visual Recognition IBM Watson visual recognition capabilities include face and visual analysis via IBM cloud services. | managed ML | 7.7/10 | 8.0/10 | 7.7/10 | 7.4/10 |
| 7 | Sightengine Sightengine delivers face detection and face-related attributes through API endpoints for media moderation workflows. | moderation API | 7.4/10 | 7.2/10 | 7.5/10 | 7.5/10 |
| 8 | Kairos Kairos provides face recognition and related identity matching services via a hosted platform. | face recognition | 7.1/10 | 6.9/10 | 7.4/10 | 7.1/10 |
| 9 | Face++ Face++ offers face detection and recognition APIs for identity and verification workflows. | verification API | 6.8/10 | 7.1/10 | 6.5/10 | 6.7/10 |
| 10 | TrueFace TrueFace provides face recognition APIs for detecting and matching faces in images and video frames. | face recognition | 6.5/10 | 6.5/10 | 6.3/10 | 6.7/10 |
Vision AI provides face detection and face-related analysis through Google Cloud APIs.
Rekognition offers face detection and face search capabilities through AWS APIs.
Azure Face exposes face detection and face recognition features via REST APIs on Microsoft Azure.
Clarifai provides face detection and face recognition models through hosted machine learning APIs.
AWS face search features for identifying faces are available through Rekognition services documented for production use.
IBM Watson visual recognition capabilities include face and visual analysis via IBM cloud services.
Sightengine delivers face detection and face-related attributes through API endpoints for media moderation workflows.
Kairos provides face recognition and related identity matching services via a hosted platform.
Face++ offers face detection and recognition APIs for identity and verification workflows.
TrueFace provides face recognition APIs for detecting and matching faces in images and video frames.
Google Cloud Vision AI
API-firstVision AI provides face detection and face-related analysis through Google Cloud APIs.
Face detection with landmarks plus emotion and safety signals in a single Vision request
Google Cloud Vision AI stands out for production-grade computer vision models delivered through managed APIs. It can detect faces, analyze emotions, and return landmark, logo, text, and safe-search results from images. The service also supports OCR with layout awareness and can process images stored in Cloud Storage. Integrations with Cloud AI Platform and event-driven workflows make it suitable for automated visual pipelines.
Pros
- Strong face detection with bounding boxes and landmark extraction
- OCR supports text detection with structure for better document parsing
- SafeSearch filters identify adult, violence, and racy content
- Batch processing via Cloud Storage enables scalable image pipelines
- Model outputs are consistent and suitable for downstream automation
Cons
- Emotion detection accuracy varies under low light and heavy occlusion
- Face analysis depends on image quality and frontal visibility
- Real-time interactive UI requires building custom front-end services
- High-volume inference needs careful quota and request management
Best For
Teams building automated face and document vision pipelines with managed APIs
Amazon Rekognition
API-firstRekognition offers face detection and face search capabilities through AWS APIs.
Liveness detection that flags likely spoofing during face verification
Amazon Rekognition stands out for delivering managed computer vision APIs that integrate directly with AWS authentication and storage services. Face detection and face search support high-throughput processing and can compare faces against either an uploaded collection or indexed references. The service includes liveness checks to reduce spoofing risk and provides confidence scores for downstream decisioning. It also supports embedding workflows through face recognition features for building consistent identity matching pipelines.
Pros
- Managed face detection with confidence scores for automation pipelines
- Face search against Rekognition face collections for identity lookup
- Liveness detection helps reduce spoofing attempts
- Integrates with AWS storage and IAM for secure access control
- High-throughput API design supports bulk image processing
Cons
- Requires dataset curation for reliable face collection matching
- Matching performance depends on image quality and capture conditions
- Limited control over model behavior compared to custom approaches
- Video workflows are less specialized than dedicated video analytics tools
- Error analysis needs additional tooling for production debugging
Best For
Teams building AWS-native face recognition workflows with managed APIs
Microsoft Azure Face
API-firstAzure Face exposes face detection and face recognition features via REST APIs on Microsoft Azure.
Face identification and verification using persisted face lists and similarity-based matching
Microsoft Azure Face stands out with dedicated face detection and analysis APIs built for Azure deployments. It supports face identification workflows and face verification by comparing detected faces against a stored face set. The service provides structured outputs for attributes like age range, gender, emotion, and landmarks, plus confidence scores for each detection. It also includes video-friendly handling for extracting face regions frame by frame within applications.
Pros
- Face detection and landmark extraction with per-face confidence scores
- Attribute analysis includes age range and emotion outputs
- Built for identity workflows using face identification and verification APIs
Cons
- Requires careful threshold tuning for recognition accuracy
- Limited to face-related signals rather than full identity management
- Strong compliance controls add integration overhead for some use cases
Best For
Applications needing face detection and attribute analysis with Azure identity features
Clarifai
API-platformClarifai provides face detection and face recognition models through hosted machine learning APIs.
Face embeddings with similarity search for identity matching in integrated workflows
Clarifai stands out for enterprise-grade computer vision built around multimodal recognition workflows and customizable models. The platform supports face-related tasks such as face detection, face recognition, and searchable face embeddings for identity and similarity matching. Clarifai also provides API access for integrating vision into apps and enables model training and fine-tuning for domain-specific accuracy. Governance features such as access controls and auditability support production deployments where data handling matters.
Pros
- Face detection and face recognition via production APIs
- Searchable face embeddings for similarity and identity matching
- Custom training to adapt accuracy to specific domains
- Enterprise governance controls for secure deployments
Cons
- Face identity workflows require careful threshold and evaluation tuning
- Best results depend on consistent labeling and dataset quality
- Implementation effort increases with custom model training
Best For
Teams integrating face recognition and similarity search into enterprise apps
AWS Amazon Face Search
enterprise APIAWS face search features for identifying faces are available through Rekognition services documented for production use.
Face collections with similarity search using CompareFaces and IndexFaces workflows
Amazon Rekognition provides face search capabilities through AWS services rather than a standalone desktop product. The solution supports indexing faces from stored images and running similarity queries to find matching identities. Developers can use the API from Rekognition to build applications for surveillance, customer lookup, and photo organization workflows. Face search is delivered as part of the Rekognition face collection feature set with programmatic control over results and thresholds.
Pros
- Face collections enable scalable indexing for similarity search queries
- Programmable search returns matches with confidence scores and bounding box data
- API-driven integration fits custom apps and automated identity workflows
- Supports batch and real-time matching patterns in production systems
Cons
- Face search quality depends heavily on input image clarity and pose
- Indexing and storage management adds engineering and operational overhead
- High-volume search can require careful tuning of thresholds and workflows
Best For
Teams building face similarity search in custom applications and pipelines
IBM Watson Visual Recognition
managed MLIBM Watson visual recognition capabilities include face and visual analysis via IBM cloud services.
Custom model training combined with face detection and emotion plus landmark extraction
IBM Watson Visual Recognition is built for extracting face-related signals from images and video frames. It can detect faces and return descriptive attributes like emotions and face landmarks. It also supports custom model training for recognizing new visual categories tied to faces and scenes.
Pros
- Face detection with structured outputs for downstream workflows
- Emotion and landmark inference for richer face analytics
- Custom visual recognition models for domain-specific face-related classes
- APIs integrate with existing services for automated processing
Cons
- Results require careful labeling and threshold tuning for accuracy
- Emotion and attribute outputs can be brittle across lighting and angles
- Identity matching is not a primary focus versus face verification products
Best For
Teams needing image face analytics and custom visual classification via API
Sightengine
moderation APISightengine delivers face detection and face-related attributes through API endpoints for media moderation workflows.
Face presence and quality scoring alongside attribute extraction in a single Sightengine API
Sightengine provides automated face analytics for images and videos with API-driven skin, face, and quality signals. It can detect faces and estimate attributes like age range, gender, emotions, and skin tone where supported by the pipeline. It also produces quality and safety indicators such as face presence and blur likelihood to support moderation and document checks. The solution is distinct for combining face detection with downstream attribute extraction in one workflow for visual moderation and user verification.
Pros
- API delivers face detection plus attribute inference in one request flow
- Quality signals help filter blurry or low-visibility face inputs
- Supports batch processing for high-volume content pipelines
- Emphasis on moderation-ready outputs for trust and safety teams
Cons
- Attribute accuracy can degrade on heavily occluded or low-light faces
- Video workflows require careful frame sampling to avoid missed faces
- Fine-grained customization needs model and threshold tuning per use case
Best For
Teams needing face analytics for moderation, onboarding, and identity workflows
Kairos
face recognitionKairos provides face recognition and related identity matching services via a hosted platform.
Configurable workflow engine with role-based task routing and stage-level execution history
Kairos stands out for its focus on turning business processes into trackable automation through configurable workflows. The solution supports data ingestion, task routing, and role-based reviews so work moves through stages with audit visibility. It also emphasizes integrations for connecting internal systems and keeping process status synchronized. Kairos is built to centralize operational execution and reporting around process outcomes.
Pros
- Configurable workflow design maps process steps to measurable outcomes
- Role-based routing supports approvals and controlled handoffs
- Audit-friendly execution history improves traceability across process stages
- Integration support keeps process status aligned with connected systems
Cons
- Workflow configuration can be time-consuming for complex, multi-team processes
- Limited public documentation makes advanced setup harder to validate
- Reporting depth may require customization for highly specific KPIs
Best For
Teams needing workflow automation with approvals and strong process traceability
Face++
verification APIFace++ offers face detection and recognition APIs for identity and verification workflows.
Face verification with similarity scoring for direct identity matching
Face++ focuses on production-grade face analysis APIs for identity and verification workflows. It provides face detection, landmark extraction, face comparison, and biometric quality checks for video and images. It also supports person and attribute-related recognition tasks that fit moderation and identity verification pipelines. Documentation and SDK options target direct integration into existing backend systems.
Pros
- High-coverage face detection and landmark extraction for images and videos
- Face verification and similarity scoring for identity matching workflows
- Quality and liveness oriented signals to reduce bad input risk
- Attribute extraction helps automate tagging and downstream processing
Cons
- Face recognition accuracy can drop with occlusion and low-light inputs
- Identity workflows require careful thresholding and dataset calibration
- API-centric design needs engineering effort for end-to-end products
Best For
Teams building face verification and biometric automation via API integration
TrueFace
face recognitionTrueFace provides face recognition APIs for detecting and matching faces in images and video frames.
Human-in-the-loop review for low-confidence face similarity decisions
TrueFace focuses on face analysis for identity and verification workflows using automated facial processing. The core capabilities include detecting faces, extracting biometric features, and comparing subjects for similarity decisions. The system supports human-in-the-loop review to handle ambiguous matches and reduce false acceptances. TrueFace is best suited for organizations that need repeatable face matching across images from operational sources.
Pros
- Face detection and feature extraction designed for consistent biometric processing
- Face similarity matching supports identity verification workflows
- Human review options help handle low-confidence matches
Cons
- Decision accuracy depends heavily on image quality and pose
- Model outputs require integration to fit existing identity systems
- Fewer configuration controls than full custom biometric pipelines
Best For
Identity verification workflows needing automated face matching with review support
How to Choose the Right Face Software
This buyer’s guide covers face detection and face-related analysis tools including Google Cloud Vision AI, Amazon Rekognition, Microsoft Azure Face, Clarifai, AWS Amazon Face Search, IBM Watson Visual Recognition, Sightengine, Kairos, Face++, and TrueFace. It focuses on how to match tool capabilities to identity verification, similarity search, moderation readiness, and workflow traceability needs. It also highlights concrete pitfalls seen across these tools like threshold tuning, low-light and occlusion sensitivity, and production integration effort for UI and pipelines.
What Is Face Software?
Face software is a set of APIs and hosted services that detect faces and return structured outputs like bounding boxes, landmarks, confidence scores, and face-related attributes such as emotion, age range, and skin tone. Many tools also support face verification and similarity matching using persisted face lists, face collections, or embedding-based search. These tools solve problems in automated identity workflows, onboarding checks, and content moderation pipelines where face presence and quality signals must be extracted at scale. Google Cloud Vision AI represents a production-ready computer vision API approach, while Microsoft Azure Face targets identity workflows using face identification and verification endpoints.
Key Features to Look For
Face software selection should be driven by the exact signals and workflow primitives each tool exposes, because face quality, identity matching, and safety outputs behave differently across vendors.
Landmarks and confidence-scored face detection
Tools that output landmarks plus per-face confidence scores support reliable downstream decisions in automated pipelines. Google Cloud Vision AI provides face detection with bounding boxes and landmark extraction, and Microsoft Azure Face returns structured outputs with confidence scores for each detected face.
Identity matching primitives with face lists or face collections
Verification and similarity search require persisted references such as persisted face lists or indexed face collections. Microsoft Azure Face supports face identification and face verification by comparing detected faces against a stored face set, and AWS Amazon Face Search enables similarity search through Rekognition face collections using IndexFaces and CompareFaces workflows.
Liveness and spoofing resistance signals
Liveness helps reduce spoofing attempts in verification flows that accept a subject based on face match quality. Amazon Rekognition includes liveness checks that flag likely spoofing during face verification, and Face++ provides biometric quality checks oriented toward verification safety.
Searchable face embeddings and similarity search
Embedding-based similarity search supports identity lookup and scalable matching across large candidate sets. Clarifai offers face embeddings with similarity search for identity matching, while AWS Amazon Face Search uses face collections to return match results with confidence scores and bounding box data.
Safety and moderation-ready face signals
Moderation workflows benefit from face presence and quality signals plus safety filtering to prevent processing unsafe or unusable inputs. Sightengine delivers face presence and quality scoring alongside attribute extraction, and Google Cloud Vision AI returns SafeSearch results for adult, violence, and racy content in the same vision request.
Workflow orchestration with traceable human and stage routing
Operational deployments often need review handling for ambiguous matches and stage-level execution history for audit trails. TrueFace includes human-in-the-loop review options for low-confidence face similarity decisions, and Kairos provides a configurable workflow engine with role-based task routing and stage-level execution history for process traceability.
How to Choose the Right Face Software
Choosing the right tool depends on whether the system needs basic detection and attributes, verification with liveness, embedding-based similarity search, or workflow orchestration with audit visibility.
Map the use case to the required outputs
If the use case needs face detection plus structured landmarks and safety signals, Google Cloud Vision AI is built around face landmarks and SafeSearch in a single Vision request. If the use case needs verification-specific spoofing protection, Amazon Rekognition provides liveness checks that flag likely spoofing during face verification.
Pick the identity workflow primitive that matches the system architecture
For systems that store and reuse reference subjects inside the platform, Microsoft Azure Face compares detected faces against a stored face set using face identification and face verification. For systems that need indexed similarity search across many images, AWS Amazon Face Search uses Rekognition face collections so IndexFaces and CompareFaces power the lookup flow.
Plan for video and low-quality input handling based on tool behavior
For video or frame-based extraction, Amazon Rekognition and Azure Face both target face detection in production workflows, while Face++ supports processing for video and image verification. For moderation pipelines, Sightengine adds face presence and blur likelihood signals to filter low-visibility face inputs, but attribute accuracy can degrade on heavily occluded or low-light faces.
Decide whether to rely on attributes or focus on identity similarity only
If the system needs attribute-rich outputs such as age range, gender, emotion, and landmarks, Microsoft Azure Face provides attribute analysis with confidence scores. If the system needs identity similarity, Clarifai focuses on searchable face embeddings for similarity matching and reduces the need for separate embedding engineering.
Add workflow controls for review, thresholds, and auditability
If ambiguous matches must be reviewed before taking action, TrueFace supports human-in-the-loop review for low-confidence face similarity decisions. If operations require role-based approvals and auditable stage history, Kairos supplies a configurable workflow engine with role-based routing and stage-level execution history, while still integrating with connected systems for process status alignment.
Who Needs Face Software?
Face software fits teams building automated vision pipelines, identity verification and similarity search systems, moderation and onboarding checks, and approval-driven operations.
Teams building automated face and document vision pipelines with managed APIs
Google Cloud Vision AI fits teams that want face detection with landmarks plus emotion and safety signals delivered through managed APIs, and it can also return OCR with layout-aware text detection for documents. This combination supports automated visual pipelines where face and document signals must be processed together.
AWS-native teams implementing verification with spoofing resistance
Amazon Rekognition is suited for AWS-native identity workflows that require face collections, face search, and verification with liveness checks. Its liveness detection flags likely spoofing, and its API design integrates with AWS authentication and storage services.
Applications that need identity verification and attribute analytics inside Azure workflows
Microsoft Azure Face is a strong fit for applications that need face identification and verification against persisted face lists plus attribute outputs like age range and emotion. Its confidence-scored detections support decision threshold tuning inside Azure identity-related systems.
Enterprise app teams that need embedding-based similarity search and governance controls
Clarifai serves teams building face recognition and similarity matching into enterprise applications using searchable face embeddings. Its customizable models and governance controls support domain-specific accuracy and secure production deployments.
Common Mistakes to Avoid
Common failure modes come from misaligned workflow primitives, underestimating low-light and occlusion sensitivity, and skipping operational controls for thresholds and review.
Using face verification without liveness checks
Verification flows that accept matches without spoofing resistance increase risk from presentation attacks, which is why Amazon Rekognition includes liveness detection. Face++ also provides biometric quality and liveness-oriented signals that support verification safety when used as part of the match decision.
Skipping threshold tuning for identity matching
Identity accuracy depends on threshold and dataset calibration, which affects tools like Microsoft Azure Face and Clarifai where recognition performance requires careful threshold evaluation. Face++ and TrueFace similarly require image-quality aware similarity thresholds to avoid false accepts and false rejects.
Overrelying on emotion attributes when face visibility is poor
Emotion detection and attribute outputs degrade under low light and heavy occlusion, which impacts Google Cloud Vision AI and IBM Watson Visual Recognition. Sightengine and other tools can provide quality or blur signals, so those signals must be used to gate attribute reliance.
Assuming face search indexing is automatic without operational overhead
Face similarity search depends on indexing and storage management, which adds engineering overhead for AWS Amazon Face Search through face collection indexing. Sightengine also needs careful frame sampling for video workflows to avoid missed faces.
How We Selected and Ranked These Tools
We evaluated each face software tool on three sub-dimensions with weights that sum to one: features at 0.40, ease of use at 0.30, and value at 0.30. The overall score is calculated as overall = 0.40 × features + 0.30 × ease of use + 0.30 × value. Google Cloud Vision AI separated itself on the features dimension because it bundles face detection with landmarks plus emotion and SafeSearch signals in a single Vision request, which reduces pipeline complexity for systems that need multiple signals together.
Frequently Asked Questions About Face Software
Which face software is best for building an automated face pipeline with managed APIs?
Google Cloud Vision AI fits automated pipelines because it delivers face detection with landmarks, logo and text signals, and safe-search results in a single managed request. Amazon Rekognition fits high-throughput automation inside AWS workflows because it combines face detection and face search with indexed collections and similarity queries.
What’s the difference between face detection and face verification in these tools?
Microsoft Azure Face supports face verification by comparing detected faces against a stored face set and returning structured confidence scores. Face++ also targets verification workflows by producing similarity scoring for face comparisons across images and video.
Which tool is strongest for face search across large image libraries?
Amazon Rekognition provides face search through face collections using IndexFaces and CompareFaces workflows. Clarifai supports searchable face embeddings so identity and similarity matching can run as part of a broader multimodal pipeline.
How do tools handle video frames for face analysis?
Microsoft Azure Face supports video-friendly handling by extracting face regions frame by frame for face identification or verification. Sightengine is designed for images and videos and returns face presence and quality signals like blur likelihood alongside attribute extraction.
Which face software includes liveness checks to reduce spoofing risk?
Amazon Rekognition includes liveness checks that flag likely spoofing during face verification. TrueFace reduces false acceptances by enabling human-in-the-loop review for ambiguous similarity decisions.
Which option is better for enterprise governance and auditability around face data?
Clarifai includes governance features such as access controls and auditability that support production deployments where data handling matters. Google Cloud Vision AI fits governed cloud workflows because it integrates into managed Google Cloud AI Platform and event-driven processing patterns.
Which tool supports training custom models tied to visual categories and face-related signals?
IBM Watson Visual Recognition supports custom model training while also providing face detection and emotion or landmark extraction. Clarifai supports customizable and fine-tuned models to improve domain-specific recognition accuracy alongside face embeddings.
Which tool is best when moderation and document checks depend on face quality signals?
Sightengine fits moderation and onboarding checks because it outputs face presence, blur likelihood, and safety or quality indicators together with face attributes. Google Cloud Vision AI also helps moderation pipelines by returning safe-search results and OCR layout-aware signals alongside face detection outputs.
How can workflow automation be added around face matching decisions and approvals?
Kairos focuses on workflow automation with role-based reviews and stage-level execution history, which suits human review gates for low-confidence matches. TrueFace complements this by using human-in-the-loop review for uncertain similarity scores before final decisions are committed.
Conclusion
After evaluating 10 technology digital media, Google Cloud Vision AI stands out as our overall top pick — it scored highest across our combined criteria of features, ease of use, and value, which is why it sits at #1 in the rankings above.
Use the comparison table and detailed reviews above to validate the fit against your own requirements before committing to a tool.
Tools reviewed
aws.amazon.comdocs.aws.amazon.comReferenced in the comparison table and product reviews above.
Keep exploring
Comparing two specific tools?
Software Alternatives
See head-to-head software comparisons with feature breakdowns, pricing, and our recommendation for each use case.
Explore software alternatives→In this category
Technology Digital Media alternatives
See side-by-side comparisons of technology digital media tools and pick the right one for your stack.
Compare technology digital media tools→FOR SOFTWARE VENDORS
Not on this list? Let’s fix that.
Our best-of pages are how many teams discover and compare tools in this space. If you think your product belongs in this lineup, we’d like to hear from you—we’ll walk you through fit and what an editorial entry looks like.
Apply for a ListingWHAT THIS INCLUDES
Where buyers compare
Readers come to these pages to shortlist software—your product shows up in that moment, not in a random sidebar.
Editorial write-up
We describe your product in our own words and check the facts before anything goes live.
On-page brand presence
You appear in the roundup the same way as other tools we cover: name, positioning, and a clear next step for readers who want to learn more.
Kept up to date
We refresh lists on a regular rhythm so the category page stays useful as products and pricing change.