Difference between revisions of "2017 Winter Project Week/Evaluate Deep Learning for binary cancer legion classification"

From NAMIC Wiki
Jump to: navigation, search
 
(16 intermediate revisions by the same user not shown)
Line 2: Line 2:
 
Image:PW-Winter2017.png|link=2017_Winter_Project_Week#Projects|[[2017_Winter_Project_Week#Projects|Projects List]]
 
Image:PW-Winter2017.png|link=2017_Winter_Project_Week#Projects|[[2017_Winter_Project_Week#Projects|Projects List]]
 
Image:Cancer_roi_Img_00001.png|Example Cancer ROI
 
Image:Cancer_roi_Img_00001.png|Example Cancer ROI
 +
Image:Cancer-roi-digits-training-0109.png|Initial LeNet training with DIGITS/Caffe
 +
Image:KVis-trained-LeNet-data-augmentation.png|Training performance after data augmentation
 +
KVis-deep-learning-project.pdf | Presentation Slides on the project
 +
KVis-deep-learning-video.mov | Project Demonstration Video
 
<!-- Use the "Upload file" link on the left and then add a line to this list like "File:MyAlgorithmScreenshot.png" -->
 
<!-- Use the "Upload file" link on the left and then add a line to this list like "File:MyAlgorithmScreenshot.png" -->
 
</gallery>
 
</gallery>
Line 7: Line 11:
 
==Key Investigators==
 
==Key Investigators==
 
<!-- Add a bulleted list of investigators and their institutions here -->
 
<!-- Add a bulleted list of investigators and their institutions here -->
−
Curt Lisle, KnowledgeVis, LLC
+
*Curt Lisle, KnowledgeVis, LLC
−
others are invited
+
*Yanling Liu, FNLCR
  
 
==Project Description==
 
==Project Description==
Line 21: Line 25:
 
|
 
|
 
<!-- Approach and Plan bullet points -->
 
<!-- Approach and Plan bullet points -->
−
* Curt has a dataset prepared by colleagues at the Frederick National Lab for Cancer Research to use to train a classifier.  
+
* Start with a dataset, prepared at the Frederick National Lab for Cancer Research, to use to train a classifier.  
−
* The library consists of a set of 50x50 PNG images containing cancer lesions and two different 50x50 sets which do not contain lesions.  
+
* The dataset consists of a series of 52x52 Region Of Interest T2 MR images containing cancer lesions and two T2 image series which do not contain lesions.  
−
* We plan to collect advice from others at the Project Week to select a deep learning framework and attempt to build a classifer using this training data.
+
* We plan to collect advice from others at the Project Week,  select a deep learning framework, and attempt to build a classifer using this training data.
 
|
 
|
 
<!-- Progress and Next steps bullet points (fill out at the end of project week) -->
 
<!-- Progress and Next steps bullet points (fill out at the end of project week) -->
−
*
+
* created a DIGITS Amazon instance using NVIDIA's marketplace image before project week
 +
* Prepared the dataset in the style of the MNIST example
 +
* Trained LeNet and AlexNet CNNs using DIGITS interface and Caffe learning framework
 +
* Data augmentation was crucial to improve results up to 83% detection accuracy for 2D case
 +
* 3D data was presented without augmentation and yielded better results than 2D alone
 +
* We believe results will further improve when better data augmentation and 3D slice data are used simultaneously
 
|}
 
|}
  
 
==Background and References==
 
==Background and References==
 
<!-- Use this space for information that may help people better understand your project, like links to papers, source code, or data -->
 
<!-- Use this space for information that may help people better understand your project, like links to papers, source code, or data -->

Latest revision as of 15:35, 13 January 2017

Home < 2017 Winter Project Week < Evaluate Deep Learning for binary cancer legion classification

Key Investigators

  • Curt Lisle, KnowledgeVis, LLC
  • Yanling Liu, FNLCR

Project Description

Objective Approach and Plan Progress and Next Steps
  • Train a neural network to become a binary classifier for the detection of cancer lesions using T2 ROI images
  • Start with a dataset, prepared at the Frederick National Lab for Cancer Research, to use to train a classifier.
  • The dataset consists of a series of 52x52 Region Of Interest T2 MR images containing cancer lesions and two T2 image series which do not contain lesions.
  • We plan to collect advice from others at the Project Week, select a deep learning framework, and attempt to build a classifer using this training data.
  • created a DIGITS Amazon instance using NVIDIA's marketplace image before project week
  • Prepared the dataset in the style of the MNIST example
  • Trained LeNet and AlexNet CNNs using DIGITS interface and Caffe learning framework
  • Data augmentation was crucial to improve results up to 83% detection accuracy for 2D case
  • 3D data was presented without augmentation and yielded better results than 2D alone
  • We believe results will further improve when better data augmentation and 3D slice data are used simultaneously

Background and References

Retrieved from "https://www.na-mic.org/w/index.php?title=2017_Winter_Project_Week/Evaluate_Deep_Learning_for_binary_cancer_legion_classification&oldid=95087"