This paper describes the submission of the BMET group to the Sub gure Classi cation and Multi-Label Classi cation tasks of the ImageCLEF 2016 medical subtrack. Our method creates a new optimised feature extractor by using medical images to ne-tune a CNN that has been pre-trained on general image data. Our classi cation method shows promising result in both the the sub gure classi cation and multi-label classi cation subtasks.
This paper describes the submission of the BMET group to two of the the
ImageCLEF 2016 Medical Tasks: Sub gure Classi cation and Multi-Label
Classi cation [
Convolutional neural networks (CNNs) were used to optimise feature extraction for the ImageCLEF Medical Tasks in 2015 [6]. In our prior work, we designed a new CNN for both modality classi cation [7] and x-ray body region identi cation [8]. Choi [9] used generic features learned by a CNN from a large, well-labelled natural image dataset. However, the size of the challenge dataset limited the ability to learn the best image features.
In this paper, we describe a method for modality classi cation that uses a smaller medical image dataset to ne-tune (optimise) a CNN that was pretrained on a large natural image dataset. This method allows us to adapt or adjust the generic features learned from natural images to be more speci c for the medical imaging modalities in the ImageCLEF datasets. We apply our method to the Sub gure Classi cation and Multi-Label Classi cation tasks.
We used the Sub gure Classi cation training dataset (6776 images, 30 classes)
to ne-tune our CNN. The test datasets consisted of 4166 images for Sub gure
Classi cation and 1084 images for Multi-Label Classi cation. A full description
of the datasets can be found in the ImageCLEF 2016 overview papers [
We used the well-established AlexNet architecture [10] pre-trained (initialised) on the ImageNet natural image dataset (1000 classes, > 1 million samples) [11]. We ne-tuned the initial AlexNet lter weights (derived from natural images) for 100 epochs using back-propagation so that they were more appropriate for the 30 classes in the dataset. Dropout was used to avoid over tting.
We increased the robustness of our algorithm to translation and orientation using a 24-fold data augmentation scheme. We generated 6 crops (original, top left, top right, bottom left, bottom right, centre) and 4 re ections (no ip, x axis, y-axis and both axes) of each training sample; 90% of the augmented dataset was used for ne-tuning and 10% for validation.
The ne-tuned CNN produced a 4096-dimensional feature vector for each input image. To improve e ciency, we reduced the dimensionality using Principle Component Analysis (PCA) [12] to select the principle components that explained 99% of the variation in the data (1453 dimensions).
We trained a multi-class support vector machine (SVM) using the PCAreduced features extracted from all 24 augmented variations of the training dataset. During classi cation, we generated the feature vectors for each test image and its 5 crops, and used the SVM to obtain the posterior probability and per-class score that each crop depicted a particular modality. When using per-class SVM scores, we linearly scaled them to the range [0; 1] to reduce the impact of very large outlier scores. We investigated several di erent schemes to determine the class of an input image, as described in our runs.
We implemented our method in MATLAB, using the MatConvNet library [13] for our implementation of CNN ne-tuning. For our experiments we used the pre-trained AlexNet provided as a part of MatConvNet. 4
We submitted 5 runs to the Sub gure Classi cation (SC) task and 2 runs to the Multi-Label Classi cation (ML) task.
SC1 Mean SVM posterior probabilities of all crops.
SC2 Mean of the per-class SVM score, which were scaled across all crops. SC3 Mean of the per-class SVM scores, scaled separately for each crop. SC4 Majority class across all crops. The per-class scores were not scaled. SC5 Maximum SVM posterior probability across all crops.
ML1 For each crop, the label was the modality with the highest SVM score. ML2 For each crop, the label was the modality with the highest probability. Acknowledgments. We gratefully acknowledge the support of NVIDIA Corporation through the donation of the Tesla K40 GPU used for this research.
This work was supported in part by ARC grants. This work was also supported by the Faculty of Engineering and Information Technologies, The University of Sydney, under the Faculty Research Cluster program. 3. Kumar, A., Kim, J., Cai, W., Feng, D.: Content-based medical image retrieval: a survey of applications to multidimensional and multimodality data. Journal of Digital Imaging 26(6) (2013) 1025{1039 4. Pelka, O., Friedrich, C.M.: FHDO biomedical computer science group at medical classication task of ImageCLEF 2015. In: CLEF 2015 Working Notes. Volume 1391. (2015) 5. Abedini, M., Cao, L., Codella, N., Connell, J.H., Garnavi, R., Geva, A., Merler, M., Nguyen, Q.B., Pankanti, S.U., Smith, J.R., Sun, X., Tzadok, A.: IBM research at ImageCLEF 2013 medical tasks. In: CLEF 2013 Working Notes. Volume 1179. (2013) 6. Villegas, M., Muller, H., Gilbert, A., Piras, L., Wang, J., Mikolajczyk, K., de Herrera, A.G.S., Bromuri, S., Amin, M.A., Mohammed, M.K., Acar, B., Uskudarli, S., Marvasti, N.B., Aldana, J.F., del Mar Roldan Garc a, M.: General Overview of ImageCLEF at the CLEF 2015 Labs. Lecture Notes in Computer Science. (2015) 7. Lyndon, D., Kumar, A., Kim, J., Leong, P.H.W., Feng, D.: Convolutional neural networks for medical classi cation. In: CLEF 2015 Working Notes. Volume 1391. (2015) 8. Lyndon, D., Kumar, A., Kim, J., Leong, P.H.W., Feng, D.: Convolutional neural networks for medical clustering. In: CLEF 2015 Working Notes. Volume 1391. (2015) 9. Choi, S.: X-ray Image Body Part Clustering using Deep Convolutional Neural Network: SNUMedinfo at ImageCLEF 2015 Medical Clustering Task. In: CLEF 2015 Working Notes. (2015) 10. Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classi cation with deep convolutional neural networks. In Pereira, F., Burges, C., Bottou, L., Weinberger, K., eds.: Advances in Neural Information Processing Systems 25. Curran Associates, Inc. (2012) 1097{1105 11. Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., Berg, A.C., Fei-Fei, L.: Imagenet large scale visual recognition challenge. International Journal of Computer Vision 115(3) (2015) 211{252 12. Jolli e, I.: Principal component analysis. Wiley Online Library (2002) 13. Vedaldi, A., Lenc, K.: MatConvNet { Convolutional Neural Networks for MAT
LAB. In: Proceeding of the ACM Int. Conf. on Multimedia. (2015) 689{692 14. He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. arXiv preprint arXiv:1512.03385 (2015)