Abstract
In the medical imaging domain, nonlinear warping has enabled pixel-by-pixel mapping of one image dataset to a reference dataset. This co-registration of data allows for robust, pixel-wise, statistical maps to be developed in the domain, leading to new insights regarding disease mechanisms. Deep learning technologies have given way to some impressive discoveries. In some applications, deep learning algorithms have surpassed the abilities of human image readers to classify data. As long as endpoints are clearly defined, and the input data volume is large enough, deep learning networks can often converge and reach prediction, classification and segmentation with success rates as high or higher than human operators. However, machine learning and deep learning algorithms are complex. Interpretability is not always a product of the classifications performed. Visualization techniques have been developed to add a layer of interpretability. The work presented here builds on a framework for augmenting statistical findings in medical imaging workflows with machine learning results. Utilizing the framework, visualization techniques for the machine learning portion are compared in an application, and a novel, lightweight technique for machine learning visualization is proposed as a means of increasing the portability of machine learning interpretability to Internet of Things applications. The novel visualization, hierarchical occlusion, can improve time to visualization by three orders of magnitude over a traditional occlusion sensitivity algorithm.
Similar content being viewed by others
References
Monroe WS, Anthony T, Tanik MM, Skidmore FM (2019) Towards a framework for validating machine learning results in medical imaging: opening the black box. In: Proceedings of the practice and experience in advanced research computing on rise of the machines (learning). ACM, p 74
Chattopadhay A, Sarkar A, Howlader P, Balasubramanian VN (2018) Grad-cam++: generalized gradient-based visual explanations for deep convolutional networks. In: 2018 IEEE winter conference on applications of computer vision (WACV). IEEE, p 839–847
Lundberg SM , Lee S-I (2017) A unified approach to interpreting model predictions. In: Advances in neural information processing systems, pp 4765–4774
Grover S, Bhartia S, Yadav A, Seeja KR et al (2018) Predicting severity of Parkinson’s disease using deep learning. Procedia Comput Sci 132:1788–1794
Zaharchuk G, Gong E, Wintermark M, Rubin D, Langlotz CP (2018) Deep learning in neuroradiology. Am J Neuroradiol 39(10):1776–1784
Lipton ZC (2016) The mythos of model interpretability. arXiv preprint arXiv:1606.03490
Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, Laak JAVD, Van Ginneken B, Sánchez CI (2017) A survey on deep learning in medical image analysis. Med Image Anal 42:60–88
Sculley D, Gary H, Daniel G, Eugene D, Todd P, Dietmar E, Vinay C, Michael Y (2014) Machine learning: the high interest credit card of technical debt. SE4ML: software engineering for machine learning (NIPS 2014 Workshop)
Selvaraju RR, Das A, Vedantam R, Cogswell M, Parikh D, Batra D (2016) Grad-cam: Why did you say that? Visual explanations from deep networks via gradient-based localization. CoRR, abs/1610.02391, 7
Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D , Batra D (2017) Grad-cam: visual explanations from deep networks via gradient-based localization. In: 2017 IEEE international conference on computer vision (ICCV). IEEE, pp 618–626
Yang C, Rangarajan A, Ranka S (2018) Visual explanations from deep 3D convolutional neural networks for alzheimer’s disease classification. arXiv preprint arXiv:1803.02544
Anthony T, Robinson JP, Marstrander JR, Brook GR, Horton M, Skidmore FM (2017) High performance/throughput computing workflow for a neuro-imaging application: provenance and approaches. In: Suh SC, Anthony T (eds) Big data and visual analytics. Springer, pp 245–256
Kraguljac NV, Anthony T, Skidmore FM, Marstrander J, Morgan CJ, Reid MA, White DM, Jindal RD, Skefos NHM, Lahti AC (2019) Micro-and macrostructural white matter integrity in never-treated and currently unmedicated patients with schizophrenia and effects of short term antipsychotic treatment. Biol Psychiatry Cogn Neurosci Neuroimaging 4:462–471
Yin J, Anthony T, Marstrander J, Liu Y, Burdyshaw C, Horton M, Crosby L, Glenn BR, Skidmore F (2016) Optimization of non-linear image registration in AFNI. In: Proceedings of the XSEDE16 conference on diversity, big data, and science at scale. ACM, p 6
Wang L, Wu Z, Karanam S, Peng K-C, Singh RV (2018) Reducing visual confusion with discriminative attention. arXiv preprint arXiv:1811.07484
Scarpato N, Pieroni A, Di Nunzio L, Fallucchi F (2017) E-health-IoT universe: a review. Management 21(44):46
García-Magariño I, Muttukrishnan R, Lloret J (2019) Human-centric ai for trustworthy iot systems with explainable multilayer perceptrons. IEEE Access 7:125562–125574
Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems, pp 1097–1105
Marek K, Jennings D, Lasch S, Siderowf A, Tanner C, Simuni T, Coffey C, Kieburtz K, Flagg E, Chowdhury S et al (2011) The parkinson progression marker initiative (ppmi). Prog Neurobiol 95(4):629–635
Arfanakis K, Fleischman DA, Grisot G, Barth CM, Varentsova A, Morris MC, Barnes LL, Bennett DA (2013) systemic inflammation in non-demented elderly human subjects: brain microstructure and cognition. PLoS ONE 8(8):e73107
Skidmore FM, Monroe WS, Hurt C, Nicholas AP, Gerstenecker A, Anthony T, Jololian L, Cutter G, Bashir A, Denny T, et al (2019) The emerging gait dysfunction phenotype in idiopathic parkinson’s disease. bioRxiv, pp 632638
Jenkinson M, Christian FB, Timothy EJB, Mark WW, Stephen MSF (2012) FSL. Neuroimage 62(2):782–790
Cox RW (1996) Afni: software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res 29(3):162–173
Pan SJ, Yang Q et al (2010) A survey on transfer learning. IEEE Trans Knowl Data Eng 22(10):1345–1359
Chollet F (2015) keras, GitHub. https://github.com/fchollet/keras
Abadi M, Barham P, Chen J, Chen Z, Davis A, Dean J, Devin M, Ghemawat S, Irving G, Isard M et al (2016) Tensorflow: a system for large-scale machine learning. OSDI 16:265–283
Kotikalapudi R (2017) keras-vis, Github. https://github.com/raghakot/keras-vis
Petsiuk V, Das A, Saenko K (2018) Rise: randomized input sampling for explanation of black-box models. arXiv preprint arXiv:1806.07421
Zeiler MD, Fergus R (2014) Visualizing and understanding convolutional networks. In :European conference on computer vision. Springer, pp 818–833
Dourado CMJM Jr, da Suane Pires P, Silva RV, da Nóbrega M, Carlos A, da Barros S, Pedro P, Filho R, de Albuquerque VHC (2019) Deep learning IoT system for online stroke detection in skull computed tomography images. Comput Netw 152:25–39
Acknowledgements
The authors would like to acknowledge Dr. U. John Tanik for making us aware of the special issue on this topic. The author’s would also like to acknowledge Thomas Anthony, for providing an initial review of the HihO results and encouraging this work.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors share no conflict of interest with the submission of this work. This manuscript represents the authors’ original work and has neither published nor has it been submitted simultaneously elsewhere.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Monroe, W.S., Skidmore, F.M., Odaibo, D.G. et al. HihO: accelerating artificial intelligence interpretability for medical imaging in IoT applications using hierarchical occlusion. Neural Comput & Applic 33, 6027–6038 (2021). https://doi.org/10.1007/s00521-020-05379-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-020-05379-4