Biomedical Visual Search and Deep Learning Workshop
Recent developments in neural networks (or deep learning) for visual recognition have attracted the interest of internet search engines and social media sites. This interest has been driven by a desire to efficiently analyze visual content in images to help generate searchable tags enabling automated classification of the images.
The biological and biomedical science communities are rife with examples of research and clinical image data that needs to be manually classified, ranging from segmenting and annotating digital pathology images to tracking masses in digital breast tomosynthesis. A large percentage of time and effort of highly trained medical professionals is spent on manual or semi-automated identification of regions of interest. There is an urgent need to develop unsupervised or minimally supervised approaches to identifying various regions of interest within biomedical images and to the classification and generation of metadata for archived images. Besides the obvious advantages of saving time and effort, such approaches can enable the development of a biomedical visual search engine which will allow researchers and clinicians to scan through large datasets and find appropriate sets of images of interest.
The purpose of this workshop is fourfold:
a. To determine the state of the art in developing hybrid and unsupervised machine learning approaches to image classification and search in biomedical sciences using deep convolutional approaches
i. Optical microscopy
ii. Digital pathology
iii. Clinical imaging
v. ‘Omics analysis
b. To determine the barriers to the application of such techniques in the biological and biomedical sciences, including:
i. Availability of training datasets and test data.
ii. Standardization of convolutional neural networks and pattern recognition approaches.
iii. Optimized utilization of available computing power.
c. To examine what we know about how humans search and classify biomedical images, and how models of human vision can inform machine learning approaches.
d. The future vision for the development of visual search engines in the biological and biomedical sciences
The topics include, but are not limited to:
- Hybrid learning approaches using minimal training data
- Pattern recognition
- Unsupervised deep learning
- Dynamic object tracking
- Multi-object classification
- Automated region annotation
- Efficient image feature searching
- Longitudinal change analysis
- Psychophysics of visual search in complex images
- Hybrid visual-memory search in humans
- Sept 1, 2015: Due date for full workshop papers submission
- Sept 30, 2015: Notification of paper acceptance to authors
- Oct 17, 2015: Camera-ready of accepted papers
- Nov 9-12, 2015: Workshops
- We call for original and unpublished research contributions to this workshop
- Please submit a full length paper (up to 6 page IEEE 2-column format) through the online submission system (you can download the format instruction here (http://www.ieee.org/conferences_events/conferences/publishing/templates…). Electronic submissions in PDF format are required.
- Online submission system is at https://wi-lab.com/cyberchair/2015/bibm15/scripts/ws_submit.php
- Vinay Pai PhD, National Institute of Biomedical Imaging and Bioengineering, NIH
- Richard Conroy PhD, National Institute of Biomedical Imaging and Bioengineering, NIH
- Todd Horowitz PhD, National Cancer Institute, NIH
- Susan Gregurick PhD, National Institute of General Medical Sciences, NIH
- Tom Radman PhD, National Institute of Drug Abuse, NIH
- Tom Radman, NIDA
- Todd Horowitz, NCI
- Susan Gregurick, NIGMS
- Aude Oliva, MIT (tentative)
- Michael Miller, JHU (tentative)