|
Project Information
|
Title: Automated Prediction of Future Breast Cancer Occurrence from Non-Cancerous Mammograms
Subject: Breast Cancer
Subcategory: Breast Cancer Prediction
Grade level: High School - Grades 9-12
Academic Level: Ordinary
Project Type: Building / Engineering
Cost: Medium
Awards: Global Finalist
Affiliation: Google Science Fair
Year: 2016
Materials: Digital mammographic breast images,
Techniques: Bag-of-words model, signal processing, image processing
Concepts: Cancer Risk Prediction Model
Description: There is an urgent need to develop a tool that can accurately predict an individual's future cancer risk. In this project an automated way to predict a patient's future cancer risk from non-cancerous screening mammograms. Instead of focusing only on the percentage of dense tissue in the breast, was used a novel bag of words approach to represent each mammogram as a set of histogram features. Then, was built an ensemble classification system to predict future cancer risk from these histogram features. This system achieved an AUC of 0.84 which is 33% higher than what was achieved before. Through further analyses, was also found that the region of the breast closest to the nipple area has the highest predictive power, and dense and non-dense tissue regions are both crucial for determining breast cancer risk.
Link: www.googlesciencefair.com...
|
|
Background
|
|
|
Wikimedia Commons
Normal (left) versus cancerous (right) mammography image.
|
Source: Wikipedia (All text is available under the terms of the Creative Commons Attribution-ShareAlike License)
|
|
Useful Links
|
Science Fair Projects Resources
Citation Guides, Style Manuals, Reference
General Safety Resources
Electrical Safety FAQ
Breast Cancer Science Fair Projects
Cancer Causes, Diagnosis and Therapy Science Fair Projects
Health, Medicine and Human Anatomy Science Fair Projects
Medicine Experiments
|
|
Books
|
|
|
