Check Me

• 

  Image analysis

• 

  Image predictions and result

• 

  Image upload

• 

  Ai driven education

Inspiration

We initiated this project in memory of my sister, who passed away from breast cancer in June 2023. This loss motivated us to develop a solution for breast cancer specifically tailored to Africa.

What It Does

Our project leverages a combination of AI and community support. The AI component drives education tailored for African women, analyzes images, and predicts with accuracy whether the sample has breast cancer. Additionally, the community support aspect connects patients with doctors and pairs women battling breast cancer with survivors

How We Built It

We built our solution by utilizing TensorFlow to train a dataset of benign and malignant breast images.

Challenges We Encountered

We faced several challenges, including:

• Dataset Identification and Training: Finding and training an appropriate dataset was time-consuming, and we required significant computing power.
• Compatibility Issues: Ensuring that Python remained compatible across our devices proved difficult.
• UI Design Decisions: We had challenges deciding on the user interface's appearance.
• TensorFlow Learning Curve: As it was our first experience with TensorFlow, we encountered issues pushing our model (.h5 file) and the dataset to the repository. We ultimately had to rewrite the git history to untrack these files.

Accomplishments We're Proud Of

We take pride in developing a model that aids doctors with AI-driven analysis.

What We Learned

• We gained insights into Convolutional Neural Networks (CNNs), TensorFlow, and the application of machine learning for analysis.
• We learned about breast cancer statistics in Africa.
• We recognized the limitations of current solutions such as mammography, ultrasound, and clinical breast exams.

What's Next for Check Me

• We are committed to engaging with doctors and partnering with medical facilities to have our device tested.
• We aim to explore the manufacturing of our device.
• We will continue refining our model, ideally incorporating Africa-centric data to enhance its accuracy.

Built With

• flask
• nextjs
• python
• tailwind
• tensorflow