Strike-a Pose - Best Practices

DIGIMUSE ENTER BEST PRACTICES

Back to Best Practices

Strike-a Pose

Overview

Objectives:

Enhance Museum Engagement through Gamification
Encourages visitors to interact with artworks in a fun and immersive way. It also uses pose-matching challenges to spark curiosity and deepen knowledge about art.
Promote Art Education in an Interactive Manner
Provides contextual information about artworks as users match poses.
Ensure Accessibility and Inclusivity
Offer different challenge modes, including torso-only poses for wheelchair users.
Facilitate Social Sharing and Digital Engagement
Enables users to generate and share videos of their participation and creates a digital extension of the museum experience through social media.
Support Museum Customization and Reusability
Allows museums to replace default artworks with their collections.
Adopt a Privacy-Conscious, BYOD Approach
Ensures no personal data is logged, aligning with privacy best practices.
Provide Easy Setup and Deployment for Institutions
Uses Docker for simplified backend installation and allows museums to manage interactions via an admin dashboard.

Target group:

•   Young people and students
•   Museum visitors
•   Tech and gamification enthusiasts

Info

Organisation name: Media Integration and Communication Center (MICC)

Italy

Activity:

Advanced research in the fields of computer vision, multimedia technologies applied to smart environments, natural interaction, Internet-based applications, and collective intelligence.

Funding sources:

The project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 101004545

Contact

marco.bertini@unifi.it

Website

Link to initiative:

reinherit-digital-hub

Strengths

User-friendly interface – Easy adoption by museums and visitors
Low start-up costs – Open-source tech eliminates licensing fees
Minimal hardware requirements – Works on mid-range devices, reducing infrastructure costs
BYOD compatibility – Visitors can use their own smartphones, reducing museum costs
Gamified engagement – Increases visitor interaction, especially among younger audiences
Customisation – Museums can tailor content and challenges to their collections and needs
Privacy-friendly AI – Client-side processing ensures privacy and compliance with GDPR

Weaknesses

Artificial vision systems and expression recognition are one of the focal points of the ethical debate on artificial intelligence and, not surprisingly, they represent one of the most closely monitored technologies within the new AI Act of the European Union. It is therefore essential for the museum, both ethically and legally, to clarify that no automatic identity recognition policy is implemented, nor is any record of the user's physical traits. The application should cover a broad range of physical traits; similarly, it is important that apps like Strike-a-Pose work properly with users with various disabilities, such as wheelchair users. In the case of Strike-a-Pose, the curator can select only some parts of the artwork to replicate, such as the bust. Finally, since this type of app uses images of artworks, it is important for the museum to properly manage the usage and sharing rights, whether the artworks belong to the museum or other museums. In line with these ethical considerations, the apps developed in the ReInHerit project are based on specific 'ReInHerit Ethics Cards,' developed by the Consortium partners to address various issues, such as the correct training of artificial intelligence, user data protection, and respect for the rights to the images of the artworks used. (20190317_museums-and-ai-toolkit_rl_web_ita_v2-1.pdf)

Digital Solutions

Programming Languages & Frameworks

• JavaScript
• Python
• HTML5

AI & Machine Learning

TensorFlowJS MoveNet – Real-time pose detection (detects 17 key body points)
Web Technologies & APIs
Flask (Python Framework) – Used for building RESTful APIs
RESTful APIs – Enables communication between the knowledge base and the interface

Database & Data Management

SQLite – Stores museum collections, challenges, and artwork metadata

Deployment & Hosting

Docker – Simplifies backend setup and deployment

Gamification & User Interaction

Pose-matching AI – Enables Strike-a-Pose to analyse and evaluate poses
User-generated video creation – Allows users to save and share their interactive experience
Admin Dashboard – Enables museums to manage artworks and user interactions

Privacy & Security

Client-side AI processing – Ensures privacy by performing pose detection on the user’s device (no personal data logging)

Cross-Platform Compatibility

Responsive Web Design – Ensures adaptability to mobile devices, tablets, and desktops
Bring Your Own Device (BYOD) Support – Allows visitors to use their own smartphones for interaction

Demonstrable positive impacts

The Strike-a-pose app won a Best Demo Honorable Mention award at ACM Multimedia 2022, the foremost conference on multimedia.

Skills & knowledge required

1. Technical Skills

Web Development
AI and Machine Learning
Database Management
App Customisation and Deployment

2. Project Management Skills

Planning & Coordination
Budgeting & Resource Allocation
Risk Management

3. Interpersonal Skills

Communication & Collaboration
Customer-Centric Mindset
Stakeholder Engagement

4. Domain-Specific Expertise

Cultural Heritage Knowledge
Gamification & User Engagement
Legal & Ethical Knowledge

5. Attitudes

Openness to Innovation
Adaptability
Commitment to Sustainability

Transferable innovative principles and methods

Strike-a-Pose: Gamified Pose-Matching with Artwork Figures
Strike-a-Pose is a web application for pose analysis, comparing human postures with those in famous paintings or statues. It enhances museum engagement by challenging users to recreate specific poses, unlocking artwork details as they progress.

• Key Features:

Can be used on visitors' smartphones (Bring Your Own Device – BYOD) or via on-site installations with a large screen and camera
Uses TensorFlowJS MoveNet, an AI model detecting 17 body key points for pose matching in real time
Runs client-side for improved privacy and performance
Supports solo or competitive challenges with multiple users
Museums can customise content, adapting challenges to different accessibility needs (e.g., wheelchair users)
Generates shareable videos of the interactive experience

This application provides low-cost, scalable digital solutions that small museums and cultural institutions can adopt to modernise visitor experiences:

1. Low Technical Barriers

Built on open-source technologies (JavaScript, Python, HTML5, Flask, SQLite)
Uses Docker for easy backend setup and deployment
Museums can substitute sample artworks with their own collections without modifying the core system

2. Minimal Hardware Requirements

Designed to run on mid-range mobile devices and standard PCs, reducing the need for expensive workstations
AI-powered pose recognition is performed client-side, minimizing server costs and ensuring data privacy (no personal data logging)

3. Gamification & Engagement

Enhances visitor interaction with museum content
Supports customisable challenges, allowing museums to tailor experiences based on themes, historical periods, or accessibility needs
Encourages social media sharing, increasing visibility and audience reach

Methodology

Strike-a-Pose is a web-based, AI-driven pose-matching application that enhances museum engagement through gamification. The methodology focuses on:

Pose matching: Visitors replicate poses from artworks (e.g., statues or paintings), guided by AI that evaluates their accuracy using TensorFlowJS MoveNet, a real-time pose detection model.
Client-side AI processing: All pose analysis occurs on the visitor’s device, ensuring privacy (no personal data storage).
Gamified structure: Participants unlock information and receive feedback as they progress through challenges.
Customisation options: Museums can adapt challenges to their collections, themes, or accessibility goals (e.g., torso-only for wheelchair users).
Digital engagement tools: Users receive shareable videos of their interaction, extending the museum visit online.
BYOD-friendly design: Optimized for mobile and tablet use, reducing infrastructure needs for the institution.

This modular, privacy-conscious methodology combines interactive storytelling, educational goals, and digital personalization.

Resources needed and start-up costs

Core Resources:

Open-source tech stack:
JavaScript / HTML5 frontend
Python (Flask) backend
SQLite for data management
Docker for simplified deployment
TensorFlowJS MoveNet – free for client-side pose detection.
Basic server or hosting setup – €10–€50/month.
Admin dashboard – included in the application.
Optional touchscreen/kiosk devices – €300–€1,000 per device if not BYOD.

Estimated Start-up Cost:

€500–€2,000 if using BYOD and existing infrastructure.
€3,000–€5,000+ if including touchscreen displays and in-house deployment support.

Possible low cost solution

Strike-a-Pose can be effectively implemented on a limited budget, especially for small museums or cultural venues:

Run entirely via BYOD: No need for kiosks—users access via a QR code or short link.
Deploy on free-tier cloud services (e.g., Heroku, Replit, or GitHub Pages for frontend).
Use freely available artworks (e.g., public domain collections) to avoid licensing issues.
Offer simplified pose challenges that focus on upper body, reducing the need for full-body tracking and accommodating more users.
Skip video-sharing features if server load or storage is a concern—offer static image capture or live feedback only.

Estimated Low-Cost Implementation: