-
Understanding Deep Neural Network Predictions for Medical Imaging Applications
Authors:
Barath Narayanan Narayanan,
Manawaduge Supun De Silva,
Russell C. Hardie,
Nathan K. Kueterman,
Redha Ali
Abstract:
Computer-aided detection has been a research area attracting great interest in the past decade. Machine learning algorithms have been utilized extensively for this application as they provide a valuable second opinion to the doctors. Despite several machine learning models being available for medical imaging applications, not many have been implemented in the real-world due to the uninterpretable…
▽ More
Computer-aided detection has been a research area attracting great interest in the past decade. Machine learning algorithms have been utilized extensively for this application as they provide a valuable second opinion to the doctors. Despite several machine learning models being available for medical imaging applications, not many have been implemented in the real-world due to the uninterpretable nature of the decisions made by the network. In this paper, we investigate the results provided by deep neural networks for the detection of malaria, diabetic retinopathy, brain tumor, and tuberculosis in different imaging modalities. We visualize the class activation mappings for all the applications in order to enhance the understanding of these networks. This type of visualization, along with the corresponding network performance metrics, would aid the data science experts in better understanding of their models as well as assisting doctors in their decision-making process.
△ Less
Submitted 19 December, 2019;
originally announced December 2019.
-
Skin Lesion Segmentation and Classification for ISIC 2018 by Combining Deep CNN and Handcrafted Features
Authors:
Redha Ali,
Russell C. Hardie,
Manawaduge Supun De Silva,
Temesguen Messay Kebede
Abstract:
This short report describes our submission to the ISIC 2018 Challenge in Skin Lesion Analysis Towards Melanoma Detection for Task1 and Task 3. This work has been accomplished by a team of researchers at the University of Dayton Signal and Image Processing Lab. Our proposed approach is computationally efficient are combines information from both deep learning and handcrafted features. For Task3, we…
▽ More
This short report describes our submission to the ISIC 2018 Challenge in Skin Lesion Analysis Towards Melanoma Detection for Task1 and Task 3. This work has been accomplished by a team of researchers at the University of Dayton Signal and Image Processing Lab. Our proposed approach is computationally efficient are combines information from both deep learning and handcrafted features. For Task3, we form a new type of image features, called hybrid features, which has stronger discrimination ability than single method features. These features are utilized as inputs to a decision-making model that is based on a multiclass Support Vector Machine (SVM) classifier. The proposed technique is evaluated on online validation databases. Our score was 0.841 with SVM classifier on the validation dataset.
△ Less
Submitted 13 August, 2019;
originally announced August 2019.
-
Skin Lesion Segmentation and Classification for ISIC 2018 Using Traditional Classifiers with Hand-Crafted Features
Authors:
Russell C. Hardie,
Redha Ali,
Manawaduge Supun De Silva,
Temesguen Messay Kebede
Abstract:
This paper provides the required description of the methods used to obtain submitted results for Task1 and Task 3 of ISIC 2018: Skin Lesion Analysis Towards Melanoma Detection. The results have been created by a team of researchers at the University of Dayton Signal and Image Processing Lab. In this submission, traditional classifiers with hand-crafted features are utilized for Task 1 and Task 3.…
▽ More
This paper provides the required description of the methods used to obtain submitted results for Task1 and Task 3 of ISIC 2018: Skin Lesion Analysis Towards Melanoma Detection. The results have been created by a team of researchers at the University of Dayton Signal and Image Processing Lab. In this submission, traditional classifiers with hand-crafted features are utilized for Task 1 and Task 3. Our team is providing additional separate submissions using deep learning methods for comparison.
△ Less
Submitted 18 July, 2018;
originally announced July 2018.
