Glad to announce iMIMIC, a new workshop on Interpretability of Machine Intelligence in Medical Image Computing at Miccai 2018.
Check out our website at https://imimic.bitbucket.io where we will be posting more information
Check out as well our promotional video here:
The Clinical Neuroscience Bern network organises every year the Science Slam event, a great initiative to bring neuroscientists to one place; share ideas and challenges in an entertaining and informative manner.
If you haven’t yet, check out our 2017 contribution to the event!
Or..and did I mention admission is free? #brainweekbern @inselgruppe @unibern
Mark your calendar March 14th 2018, brainweekbern.ch
Whether you are teacher, student, computer scientist, or proficient machine learning programmer, there are many times where having a solid reference library on the topic can save you a lot of time and help you to prepare material for your next lecture, article, even job interview. Machine learning algorithms and lately, deep learning, have in fact demonstrated excellent results and produced many breakthroughs in computer science. This is revolutionising many fields, including healthcare where medical records, medical images, and other patient-specific information are combined with advanced machine learning approaches to create advanced algorithms capable of performing high levels of data mining and leveraging patient treatment. That being said, there is a lot of hype about Machine Learning, Artificial Intelligence (A.I), wrong expectations about what A.I is, and what it can actually do. While the Internet is full with resources about the topic, there is nothing better than learning from leading researchers and educators in the topic. This motivated me to create a list of recommended books from different authors, who also provide a different view and focus to the topic, as well as describing challenges and future opportunities.
So here it is, my list of top seven books about machine/deep learning that I’d recommend you to definitely check out. This list includes general purpose books about machine and deep learning, as well as more specific ones focusing on machine and deep learning for medical imaging. Some very good ones focusing on implementational aspects with Python and Tensorflow, for those readers who look for practical examples and more hands-on focused learning.
I hope you like. Feel free to use the links below to know more details, and happy reading!
|Books that focus on Machine Learning and Deep Learning
|Deep Learning with Python by Francois Chollet
Fresh from the oven, this has been an expected book since the first chapters were made available for free online. The book is excellent and Francois Chollet did a great job at explaining difficult concepts.
|Python Machine Learning: Machine Learning and Deep Learning with Python, scikit-learn, and TensorFlow, 2nd Edition by Sebastian Raschka, Vahid Mirjalili
On its second, revised and improved edition, this book is an excellent teaching material, guiding the reader from basic to advanced topics using main Python libraries and TensorFlow. The book features a great balance between theory and practice, and it also useful to those working on industrial applications.
|Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems by Aurélien Géron
This is hands down (almost pun intended) one of the best books out there if you want to learn by doing. Highly recommended book as it includes theory and practice, including many examples.
|Deep Learning (Adaptive Computation and Machine Learning series) by Ian Goodfellow, Yoshua Bengio, Aaron Courville
Definitely a reference book in the topic of Deep Learning. Compared to the other books listed here, this one puts more emphasis on the maths than on the practical aspects. An excellent accompanying book to the more practically-oriented ones.
|Books that focus on Machine Learning and Medical Image Computing and Analysis
|Deep Learning for Medical Image Analysis by S. Kevin Zhou, Hayit Greenspan, Dinggang Shen (Editors)
This is one of the first books focusing on theory and applications of deep learning for medical image computing. Its seventeen chapters, divided in five parts, describes state of the art approaches developed in medical image computing to solve problems dealing with object recognition, image segmentation, image parsing, image registration and synthesis, etc. Applications include a vast variety of image modalities, featuring the flexibility and power of deep learning techniques for medical image analysis. This book is specially oriented to medical image computing scientists who are entering the field of deep learning. Specially, the introductory chapters I and II are specially oriented to give the reader an introduction to neural networks and deep convolutional neural nets for computer vision.
|Medical Image Recognition, Segmentation and Parsing: Machine Learning and Multiple Object Approaches (The Elsevier and Miccai Society Book Series) 1st Edition by S. Kevin Zhou
This book from Kevin Zhou nicely complements the ones listed above with methods and approaches for image parsing and recognition. It also includes techniques and methodologies developed outside the field of deep learning, giving the reader a different view that complements the more DL-focused books listed above, and hopefully motivates the reader to consider how previous concepts and ideas could be now complemented, integrated or adapted to work with modern machine learning technologies.
|Machine Learning and Medical Imaging (Elsevier and Miccai Society) by Guorong Wu, Dinggang Shen, Mert Sabuncu (Editors)
Also from the Medical Image Computing and Computer Assisted Community (MICCAI), this book nicely presents state of the art approaches in machine learning, going from classic machine learning approaches to fundamentals of deep learning. On a second part, the book presents a plethora of applications featuring different anatomical regions and image modalities, including even applications where genetic information is used. The book definitely provides a good overview of challenges and opportunities that machine learning has for medical imaging.
Sharing recent findings on Enhancing interpretability of automatically extracted machine learning features: application to a RBM-Random Forest system on brain lesion segmentation https://authors.elsevier.com/c/1WKpf4rfPluH2D
Machine learning systems are achieving better performances at the cost of becoming increasingly complex. However, because of that, they become less interpretable, which may cause some distrust by the end-user of the system. This is especially important as these systems are pervasively being introduced to critical domains, such as the medical field. Representation Learning techniques are general methods for automatic feature computation. Nevertheless, these techniques are regarded as uninterpretable “black boxes”. In this paper, we propose a methodology to enhance the interpretability of automatically extracted machine learning features. The proposed system is composed of a Restricted Boltzmann Machine for unsupervised feature learning, and a Random Forest classifier, which are combined to jointly consider existing correlations between imaging data, features, and target variables. We define two levels of interpretation: global and local. The former is devoted to understanding if the system learned the relevant relations in the data correctly, while the later is focused on predictions performed on a voxel- and patient-level. In addition, we propose a novel feature importance strategy that considers both imaging data and target variables, and we demonstrate the ability of the approach to leverage the interpretability of the obtained representation for the task at hand. We evaluated the proposed methodology in brain tumor segmentation and penumbra estimation in ischemic stroke lesions. We show the ability of the proposed methodology to unveil information regarding relationships between imaging modalities and extracted features and their usefulness for the task at hand. In both clinical scenarios, we demonstrate that the proposed methodology enhances the interpretability of automatically learned features, highlighting specific learning patterns that resemble how an expert extracts relevant data from medical images.
Discover our new findings on automated volumetry for extent-of-resection and residual-tumor for GBM patients. Journal of NeuroSurgery Meier et al. 2016.
New findings in #BrainTumors volumetry. BraTumIA @unibern @inselgruppe and @BrainLab SmartBrush. Multiple experts http://dx.doi.org/10.1371/journal.pone.0165302
New developments in stroke tissue estimation. Introducing FASTER – Fully Automated Stroke Tissue Estimation using Random forest classifiers. McKinley et al. JCBFM 2016 http://dx.doi.org/10.1177/0271678X16674221
Highly grateful for recognition @miccai2016 Collaborative effort @unibern @inselgruppe Stefan Bauer Young Scientist Publication Impact Award
Interested in advances in Glioma, MS, Stroke, and trauma brain injuries? The Brain Lesions Miccai Workshop is happening again this year! http://www.brainlesion-workshop.org
New in Plos One – FISICO: Fast Image SegmentatIon COrrection
And!…the software tool is available for download here