Inspiration

As someone who has lived and worked closely with communities across rural Central and East Africa, I’ve witnessed firsthand the devastating consequences of delayed or inaccurate disease diagnosis. In many villages, people walk hours to the nearest clinic, only to find that basic diagnostic tools are unavailable. Diseases like malaria, tuberculosis, and anemia go undetected, leading to preventable suffering and death.

This inspired me to ask: What if we could put powerful, intelligent diagnostic tools in the hands of every community health worker, without needing a lab, internet, or electricity?

That question became the foundation of BioScan AI.

What it does

BioScan AI is a portable, AI-powered diagnostic tool designed for rural African health workers. Using just a smartphone, a clip-on lens, and our mobile app, it enables rapid, on-the-spot screening for common but deadly diseases, without internet, a lab, or advanced training.

Key Features:

Offline AI Diagnostics: Runs disease detection models on-device using images and sounds, no internet required. Disease Screening: Detects signs of malaria, anemia, and tuberculosis using: Blood smear images via the smartphone camera Nail and eye coloration for anemia Cough sounds and symptoms for TB Voice-Guided Interface: Uses local language audio prompts to guide users through each step, perfect for low-literacy settings. Instant Results: Delivers results in minutes, with risk levels and next steps. Local Data Storage: Stores diagnostic results securely on the device for reporting and follow-up.

How we built it

We began by identifying three core conditions with high impact and diagnostic potential: malaria, tuberculosis, and anemia. We gathered open-source health datasets and began designing AI models to detect:

Malaria via blood smear image analysis Anemia through nail/eye color detection TB via cough audio and symptom screening

We also designed a modular mobile app using TensorFlow Lite for offline AI, paired with a clip-on microscope lens that transforms any smartphone into a diagnostic device. The app guides health workers through the process using voice prompts in local languages, and stores all results locally.

So far, we’ve developed early concepts and data pipelines, and designed the technical roadmap. We are now ready to move into active prototype development.

Challenges we ran into

Data scarcity: Public datasets for African diagnostic contexts are limited. We had to find creative ways to source and simulate diagnostic images and audio.

Edge AI limitations: Running accurate models on low-power smartphones required extensive optimization and testing.

Designing for low-literacy users: Voice-guided, icon-based interfaces needed careful co-design with health workers.

Hardware compatibility: Ensuring universal clip-on functionality across a range of phone types was a tough but solvable challenge.

Accomplishments that we're proud of

Despite being in the early stages of development, BioScan AI has achieved several key milestones that lay the foundation for impact and scalability:

1. Designed a Feasible Offline AI Architecture We developed a clear technical roadmap using lightweight, edge-based AI models that run fully offline on low-cost Android devices, making it suitable for remote, under-resourced environments.
2. Identified Three High-Impact Diagnostic Use Cases Through consultations with rural health workers and medical advisors, we focused on malaria, tuberculosis, and anemia, three diseases that are common, deadly, and diagnosable with image/audio data.
3. Built Strong Health and Technical Partnerships We’ve initiated partnerships with local clinics, health NGOs, and AI engineers who are committed to co-developing and field-testing BioScan AI in real-world African settings.
4. Created Mockups and a Prototype Plan We designed user-centric mockups and a detailed 6-month prototype roadmap that addresses usability in low-literacy contexts, local language support, and minimal training requirements.
5. Selected for Submission to Africa Deep Tech Challenge Being selected to submit to the ADTC is itself a milestone, a validation of our idea’s potential to transform rural healthcare and improve access to diagnostics across the continent.

What we learned

Through extensive research, conversations with health workers, and early collaboration with local clinics and technologists, we learned several key insights:

Lab access is a bottleneck: Over 60% of rural clinics lack diagnostic infrastructure. Health workers are motivated but under-equipped: They need tools that are simple, portable, and multilingual. AI can work offline: With edge computing, we can run diagnostic models on low-cost smartphones without internet access. Design must be local-first: UX must work for low-literacy users, in local languages, and withstand harsh environments.

What's next for BioScan AI - Portable Disease Detection for All

With a clear vision and technical plan in place, our next step is to bring BioScan AI from concept to functional prototype and field validation. Winning the Africa Deep Tech Challenge would give us the critical resources to accelerate development and prove real-world impact.

Phase 1: MVP Development (Next 3 Months) Build and optimize lightweight AI models for malaria, anemia, and TB detection Develop the offline Android app with local-language voice support Integrate clip-on lens hardware and test image acquisition workflows Conduct internal testing with synthetic and small-scale real datasets

Phase 2: Field Pilot (Month 4–6) Deploy 10–15 BioScan AI units to rural clinics in partnership with local health NGOs Train community health workers (CHWs) to use the tool in real conditions Collect diagnostic performance data and user feedback Refine UX, voice prompts, and model accuracy based on pilot results

Phase 3: Scale-Up Planning & Seed Round (Post-Month 6) Use pilot results to apply for global health tech grants, incubators, or investor funding Expand diagnostic support to more diseases (e.g., pneumonia, dengue, HIV pre-screening) Translate the app into additional African languages Begin manufacturing low-cost, durable lens kits at scale

Long-Term Vision

We envision BioScan AI in the hands of every rural health worker across Africa, a tool as essential as the stethoscope, enabling millions of timely, AI-powered diagnoses each year. With the right support, we can close the diagnostic gap and make universal health access a reality from the village up.

Built With

• android-studio
• aws-s3-/-google-drive
• bash-/-cli
• csv-/-json
• custom-wav/mp3-audio
• edge-impulse
• google-colab-/-kaggle-notebooks
• google-text-to-speech-(tts)
• java-/-kotlin
• javascript-(node.js)
• numpy-/-pandas-/-scikit-learn
• onnx
• opencv
• python
• pytorch
• sqlite
• tensorflow-lite