Deep Learning Based Lung Cancer Detection and Classification
importance 9/10
partially solved
AI in cancer detection oncology imaging stage: research
Problem
Lung cancer must be detected early from CT scans, since late diagnosis sharply reduces survival rates.
Why it matters
Lung cancer is a leading cause of mortality worldwide and early diagnosis raises survival.
What the AI did
A densely connected convolutional neural network (DenseNet) combined with adaptive boosting was trained to classify CT lung images as normal or malignant.
Results
90.85% accuracy on a 201-image CT dataset (85% train / 15% test)
Impact
Demonstrated feasibility on a small dataset; clinical applicability not established.
Record
| Venue | IOP Conference Series Materials Science and Engineering |
|---|---|
| Published | 2020 |
| Work type | conference-paper |
| Authors | N. Kalaivani, Natraj Manimaran, Dr. S Sophia, D D Devi |
| Citations | 105 (OpenAlex, at import time) |
| References | 5 |
| Model or method | DenseNet with adaptive boosting |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1088/1757-899x/994/1/012026 |
| Open access | diamond |
| Keywords | lung cancer, CT imaging, deep learning, DenseNet, AdaBoost, classification |
Importance rating
9
major — millions of lives, an industry or a field
Abstract states lung cancer is a leading worldwide cause of mortality and early diagnosis raises survival from 14% to 49%.