Abstract
Medical image classification and concept detection are two important tasks for efficient and robust medical retrieval systems and also help with downstream tasks such as knowledge discovery, medical report generation, medical question answering, and clinical decision making. We investigate the effectiveness of transfer learning on the modality classification task using state-of-the-art deep convolutional neural networks pretrained on generic images. We also compare the performance of the traditional pipeline of handcrafted features with multi-label learning algorithms with end-to-end deep learning features for the concept detection task. Experimental results on the modality classification task show that transfer learning can leverage the patterns learned from large training data to the medical domain where little labeled data is available. Moreover, results on the concept detection task show that the deep learning approach provides better and more powerful feature representations compared to handcrafted feature extraction methods. The results on both tasks suggest that deep transfer learning methods are effective in the medical domain where data is scarce.
| Language | English |
|---|---|
| Title of host publication | 2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018 |
| Place of Publication | Piscataway, NJ |
| Publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| Pages | 1-9 |
| Number of pages | 9 |
| Volume | 2018-November |
| ISBN (Electronic) | 9781728101255 |
| ISBN (Print) | 9781728101262 |
| DOIs | |
| Publication status | Published - 2018 |
| Event | 2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018 - Auckland, New Zealand Duration: 19 Nov 2018 → 21 Nov 2018 |
Conference
| Conference | 2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018 |
|---|---|
| Country | New Zealand |
| City | Auckland |
| Period | 19/11/18 → 21/11/18 |
Fingerprint
Keywords
- Concept Detection
- Convolutional Neural Networks
- Deep Learning
- Medical Image Analysis
- Medical Image Classification
- Modality Classification
- Multi-Label learning
Cite this
}
Modality classification and concept detection in medical images using deep transfer learning. / Singh, Sonit; Ho-Shon, Kevin; Karimi, Sarvnaz; Hamey, Len.
2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018. Vol. 2018-November Piscataway, NJ : Institute of Electrical and Electronics Engineers (IEEE), 2018. p. 1-9 8634803.Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › Research › peer-review
TY - GEN
T1 - Modality classification and concept detection in medical images using deep transfer learning
AU - Singh, Sonit
AU - Ho-Shon, Kevin
AU - Karimi, Sarvnaz
AU - Hamey, Len
PY - 2018
Y1 - 2018
N2 - Medical image classification and concept detection are two important tasks for efficient and robust medical retrieval systems and also help with downstream tasks such as knowledge discovery, medical report generation, medical question answering, and clinical decision making. We investigate the effectiveness of transfer learning on the modality classification task using state-of-the-art deep convolutional neural networks pretrained on generic images. We also compare the performance of the traditional pipeline of handcrafted features with multi-label learning algorithms with end-to-end deep learning features for the concept detection task. Experimental results on the modality classification task show that transfer learning can leverage the patterns learned from large training data to the medical domain where little labeled data is available. Moreover, results on the concept detection task show that the deep learning approach provides better and more powerful feature representations compared to handcrafted feature extraction methods. The results on both tasks suggest that deep transfer learning methods are effective in the medical domain where data is scarce.
AB - Medical image classification and concept detection are two important tasks for efficient and robust medical retrieval systems and also help with downstream tasks such as knowledge discovery, medical report generation, medical question answering, and clinical decision making. We investigate the effectiveness of transfer learning on the modality classification task using state-of-the-art deep convolutional neural networks pretrained on generic images. We also compare the performance of the traditional pipeline of handcrafted features with multi-label learning algorithms with end-to-end deep learning features for the concept detection task. Experimental results on the modality classification task show that transfer learning can leverage the patterns learned from large training data to the medical domain where little labeled data is available. Moreover, results on the concept detection task show that the deep learning approach provides better and more powerful feature representations compared to handcrafted feature extraction methods. The results on both tasks suggest that deep transfer learning methods are effective in the medical domain where data is scarce.
KW - Concept Detection
KW - Convolutional Neural Networks
KW - Deep Learning
KW - Medical Image Analysis
KW - Medical Image Classification
KW - Modality Classification
KW - Multi-Label learning
UR - http://www.scopus.com/inward/record.url?scp=85062770991&partnerID=8YFLogxK
U2 - 10.1109/IVCNZ.2018.8634803
DO - 10.1109/IVCNZ.2018.8634803
M3 - Conference proceeding contribution
SN - 9781728101262
VL - 2018-November
SP - 1
EP - 9
BT - 2018 International Conference on Image and Vision Computing New Zealand, IVCNZ 2018
PB - Institute of Electrical and Electronics Engineers (IEEE)
CY - Piscataway, NJ
ER -