CerviScan

Inspiration

CerviScan was inspired by conversations with women, including my wife who described how repeated internal examinations during labor can be uncomfortable and emotionally stressful. These checks are necessary for monitoring labor progress, but the process can feel invasive, especially when repeated many times over long hours.

This led to a key question:

Can we use technology to obtain the same medical information in a safer, less invasive way?

Seeing how ultrasound and AI are transforming many areas of healthcare, we explored whether external imaging combined with AI measurement could provide doctors with accurate data while improving patient comfort and dignity.

CerviScan is built around patient comfort, clinical accuracy, and better decision-making during childbirth.

What it does

CerviScan is an AI-powered prototype that measures cervical dilation using external ultrasound imaging, reducing the need for repeated internal examinations.

The system:

Captures ultrasound images using an external probe

Uses AI to identify and measure cervical dilation

Displays results instantly on a tablet or computer

Tracks labor progression over time

Provides objective measurements for clinicians

The result is a faster, repeatable, and more comfortable monitoring method.

How we built it

During the hackathon, we developed a proof-of-concept system using:

Ultrasound imaging data for simulation and testing

Image preprocessing to enhance ultrasound quality

AI segmentation models to detect the dilation opening

Algorithms that convert image measurements into centimeters

A simple dashboard to display results and progression trends

The system processes ultrasound images, identifies the dilation region, and automatically calculates measurements for clinicians.

Challenges we ran into

Key challenges included:

Limited access to medical ultrasound datasets

Variations in ultrasound image quality and angles

Ensuring measurement consistency across scans

Designing a workflow that fits real hospital environments

Balancing speed of development with clinical realism

Healthcare solutions demand accuracy and reliability, making rapid prototyping more complex.

Accomplishments that we're proud of

During the hackathon, we successfully:

Built an end-to-end prototype measuring dilation from ultrasound images

Demonstrated that external imaging can support labor monitoring

Created an AI-assisted workflow focused on patient comfort

Delivered a functional demo within limited time

Established groundwork for a scalable medical innovation

Most importantly, we proved the concept is technically possible.

What we learned

We learned that solving healthcare problems requires combining:

Medical understanding

AI and engineering skills

Human-centered design

Practical clinical workflow thinking

We also learned that even small improvements in patient experience can have meaningful impact during childbirth.

What's next for CerviScan

Next steps include:

Improving AI accuracy with larger datasets

Building real-time scanning capability

Conducting clinical pilot testing

Refining usability for nurses and doctors

Integrating with hospital monitoring systems

Preparing for clinical validation and regulatory pathways

Our long-term goal is to make labor monitoring safer, more comfortable, and accessible globally.

Built With

• browser-apis
• custom-css-animations
• es6-modules-(esm)
• gemini-3-flash-preview-model
• generative-ai-sdk
• google-gemini-api-(@google/genai)
• json
• local-state-management
• lucide-react
• react-19
• react-hooks
• recharts
• responsive-web-design
• svg-graphics
• tailwind-css
• typescript-(es6+)