Deep anomaly detection for medical images
Deep learning methods have been demonstrated to be effective in many medical tasks. However, these methods normally require a large number of labeled data, which is costly especially for disease screening, where the abnormal/diseased data are more difficult to obtain. The main purpose of this Fi...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/140534 |
| _version_ | 1811678355884343296 |
|---|---|
| author | Li, Xintong |
| author2 | Lin Zhiping |
| author_facet | Lin Zhiping Li, Xintong |
| author_sort | Li, Xintong |
| collection | NTU |
| description | Deep learning methods have been demonstrated to be effective in many medical tasks.
However, these methods normally require a large number of labeled data, which is
costly especially for disease screening, where the abnormal/diseased data are more
difficult to obtain.
The main purpose of this Final Year Project is to investigate transfer learning-based
anomaly detection methods that do not require or only require a small number of labeled
data instances. In this work, two anomaly detection methods are proposed. A
semi-supervised joint learning method (SmSupJL) is proposed to train a feature extractor
with two losses, namely, ’cross-entropy loss’ and ’intra-class variance loss’ on
a small labeled train set. By applying these two losses, the feature extractor is able
to learn discriminative features of normal and abnormal samples and keep the compactness
of normal samples. To further reduce the number of labeled data instances
needed, we propose an unsupervised domain adaptation method (UnSupDA) which
does not require any labeled instances from target domain but a small number of labeled
data instances from source domain to detect anomalies. Self-supervised tasks
are used to align source domain and target domain and thus transfer the knowledge
learned from the source domain to target domain.
Experimental results evaluated on Kaggle Diabetic Retinopathy (DR) dataset demonstrated
that the performance of these methods is either surpass or comparable to the
current state-of-the-art. |
| first_indexed | 2024-10-01T02:51:57Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/140534 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T02:51:57Z |
| publishDate | 2020 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1405342023-07-07T18:46:47Z Deep anomaly detection for medical images Li, Xintong Lin Zhiping School of Electrical and Electronic Engineering I2R ezplin@ntu.edu.sg Engineering::Electrical and electronic engineering Deep learning methods have been demonstrated to be effective in many medical tasks. However, these methods normally require a large number of labeled data, which is costly especially for disease screening, where the abnormal/diseased data are more difficult to obtain. The main purpose of this Final Year Project is to investigate transfer learning-based anomaly detection methods that do not require or only require a small number of labeled data instances. In this work, two anomaly detection methods are proposed. A semi-supervised joint learning method (SmSupJL) is proposed to train a feature extractor with two losses, namely, ’cross-entropy loss’ and ’intra-class variance loss’ on a small labeled train set. By applying these two losses, the feature extractor is able to learn discriminative features of normal and abnormal samples and keep the compactness of normal samples. To further reduce the number of labeled data instances needed, we propose an unsupervised domain adaptation method (UnSupDA) which does not require any labeled instances from target domain but a small number of labeled data instances from source domain to detect anomalies. Self-supervised tasks are used to align source domain and target domain and thus transfer the knowledge learned from the source domain to target domain. Experimental results evaluated on Kaggle Diabetic Retinopathy (DR) dataset demonstrated that the performance of these methods is either surpass or comparable to the current state-of-the-art. Bachelor of Engineering (Information Engineering and Media) 2020-05-30T09:05:33Z 2020-05-30T09:05:33Z 2020 Final Year Project (FYP) https://hdl.handle.net/10356/140534 en B3139-191 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Electrical and electronic engineering Li, Xintong Deep anomaly detection for medical images |
| title | Deep anomaly detection for medical images |
| title_full | Deep anomaly detection for medical images |
| title_fullStr | Deep anomaly detection for medical images |
| title_full_unstemmed | Deep anomaly detection for medical images |
| title_short | Deep anomaly detection for medical images |
| title_sort | deep anomaly detection for medical images |
| topic | Engineering::Electrical and electronic engineering |
| url | https://hdl.handle.net/10356/140534 |
| work_keys_str_mv | AT lixintong deepanomalydetectionformedicalimages |