An AI-driven system enhancing workplace safety by automating mask compliance monitoring post-COVID.
Role
UX/UI Designer & Web Developer
Scope
8 Months Capstone Project
Team of 4
Tools
Figma
Balsamiq
Django
A Mask Meltdown...
Why won't anyone wear them !?
Despite government mandates for mask-wearing in public, widespread non-compliance persists post-pandemic, particularly in workplaces where close proximity raises COVID-19 transmission risks. Ensuring consistent mask usage is crucial for containment, yet manual monitoring strains administrative staff, necessitating an efficient solution.
Target Audience
Age Group:
18-60 years old
Work profile:
Veritek Engineering Pvt. Ltd. Employees
Gathering Perspectives
Acknowledging the busy schedules of employees, I opted to conduct user surveys to streamline the data collection process. Administered to 42 employees across departments, the surveys aimed to efficiently gather both quantitative and qualitative insights on mask compliance.
By exploring employees' experiences and attitudes towards mask-wearing, and perceptions of existing protocols, the surveys provided valuable insights into workplace practices.
The Important Questions...
-
Do your coworkers consistently wear masks correctly?
-
How critical is mask-wearing perceived?
-
Have any of your coworkers contracted COVID-19 due to other not wearing masks?
-
How often is masks compliance checked?
Unlocking Frustrations
-
Inconsistent mask-wearing adherence.
-
Ignorance/no motivation among employees.
-
Absence of effective monitoring system.
Issue Identification
Goals
-
Automate mask detection for real-time monitoring.
-
Enable user self-monitoring and feedback.
-
Improve administrative efficiency and enhance safety.
1. Inconsistent mask-wearing adherence
Only 10 out of 42 surveyed employees reported that their coworkers consistently wear masks incorrectly, despite government mandates.
This indicates a failure in enforcement or accountability within the workplace. While mask-wearing is legally required, the absence of regular compliance checks suggests a lack of effective monitoring mechanisms. Without strict enforcement and consequences for non-compliance, employees perceive mask-wearing as optional rather than mandatory.
2. Ignorance among employees:
Despite government mandates, only 15 out of 42 employees expressed a high level of importance regarding mask-wearing.
This ignorance stems from a sense of complacency due to the easing of pandemic-related restrictions or a belief that the risk of COVID-19 transmission has diminished. Despite government mandates, the absence of immediate consequences for non-compliance may contribute to employees disregarding safety protocols.
3. Absence of effective monitoring system:
30 out of 42 respondents reported infrequent or nonexistent mask compliance checks, despite government mandates.
The lack of regular compliance checks indicates a failure in implementing and enforcing mask mandates at the organizational level. The absence of strict oversight mechanisms allows non-compliance to persist unchecked, highlighting the need for more robust monitoring and enforcement strategies.
Success Metrics
-
Compliance Rate: Increase in correct mask usage.
-
Reduction in Administrative Burden: Decrease in monitoring time.
-
User Engagement: High interaction with the platform.
-
Positive Feedback: Affirmative responses on usability and effectiveness.
Problem Deep Dive
Empathy Map
Following the user surveys, my team and I developed an empathy map to capture users' thoughts, feelings, and behaviors, to guide a user-centric design for the WorkSafe system.
This research phase informed every aspect of the design, ensuring alignment with user needs and workplace realities.
Ideation Process
System Dilemma
We struggled in determining the appropriate system to implement. Should we utilize CCTV cameras within the office? However, this method wouldn't ensure constant monitoring of individuals. Alternatively, could we develop a mobile app for penalty enforcement? Yet, there was concern that people would simply disregard it.
A Unique Approach
After multiple brainstorming sessions, we devised a solution: integrating the AI model into each employee's work device. This approach facilitates continuous monitoring as employees cannot look away and remove their masks. Additionally, the penalty system is integrated into these devices, allowing employees to receive warnings and make penalty payments before resuming work in case of three offenses. All activity logs are forwarded to the office administrator for monitoring purposes.
Finding the Best Fit
Further, we explored various technologies to determine the best web platform, AI model, and log management system for "Work Safe", including AI models compatible with Django and similar frameworks like TensorFlow, PyTorch, and OpenCV. It was challenging to find the optimal combination of technologies for seamless integration. Discussions focused on accurately issuing warnings, ensuring user-friendliness, and maximizing efficiency in enforcing mask compliance.
Technological Constraints
-
Hardware: Cameras and internet connection required.
-
Software: Compatibility and AI model accuracy crucial.
-
Data Privacy: Protection measures essential.
Redefining the Problem
In our organization, despite government mandates, there's inconsistent mask-wearing adherence, employee ignorance, and an absence of effective monitoring, leading to health risks.
User Journey Map
After deciding on the kind of AI system we wanted to build, my team and I created the user journey map.
This visual guide was crucial for identifying user interaction points with the system, streamlining the enforcement process while maintaining user engagement and compliance.
How Might We....
ensure strict mask-wearing adherence, automate monitoring, and enhance workplace safety efficiently?
Wireframes
System Implementation
Framework Selection and Model Hosting
For the implementation of the Work Safe website, Django, a high-level Python Web framework, was chosen for its efficiency in rapid development and clean design principles. Once the AI model was thoroughly tested, it was hosted on a local server.
URL routes for different pages were properly defined, and views.py managed the web requests and responses along with the underlying logic. Camera.py contained the code to import the trained models seamlessly. Additionally, a database was set up to store user data for sign-in/sign-up purposes.
Website Functionality and User Experience
Upon logging in, users are directed to the homepage where the webcam activates, capturing the user's input. The prediction is then displayed on the frame, featuring either a green box (if a mask is detected) or a red box (if not). This setup ensures a streamlined and effective experience for users navigating the website.
The Final System
Keeping An Eye
The objective is to establish a system that autonomously ensures mask compliance in the workplace, offering real-time monitoring of employees. This will alleviate the administrative staff from the manual task of enforcing mask-wearing rules, fostering a safer and more efficient work environment.
Challenges Faced
Selecting the Right System
Challenges arose due to the dilemma between CCTV cameras and a mobile app for penalty enforcement, balancing effectiveness with user acceptance.
User Experience on the Website
Ensuring a seamless and intuitive interface posed challenges, especially regarding real-time monitoring features and penalty enforcement mechanisms.
Integration of AI Model and Website
Difficulty in seamlessly integrating the AI model into the website while maintaining performance and user experience standards was a significant challenge.
User Acceptance and Compliance
Overcoming resistance to change and ensuring user compliance with the new system proved challenging, requiring effective communication and user engagement strategies.
Impact and Outcome
The deployment of the WorkSafe system at Veritek Engineering's office has yielded remarkable results. Despite being implemented in just one location, it has led to a notable improvement in mask compliance, with adherence rates increasing by over 30%.
Additionally, the manual monitoring workload has been reduced by 50%, significantly enhancing operational efficiency. Users at Veritek Engineering have praised the system's user-friendly interface and Immediate feedback, highlighting its effectiveness in promoting workplace safety.
Let's look back...
Throughout this project, my team and I encountered numerous challenges, including navigating system selection, designing user experience etc. To address usability concerns, we dedicated ourselves to refining the website interface, striving for a user-friendly experience.
Despite technical hurdles in AI integration, we remained resilient, leveraging thorough testing and optimization to overcome obstacles. This journey of perseverance and collaboration emphasized the importance of prioritizing user needs while ensuring the effectiveness of our solution.