Deep learning for biomedical image analysis

Project Title: Deep learning for biomedical image analysis
Program:
CAREERS (323)

Project Image: https://support.access-ci.org/system/files/webform/project/95/image.png
Tags:
ai (271), big-data (4), deep-learning (303), gpu (80), machine-learning (272), matlab (2), python (69), tensorflow (51)

Status: Complete
Project Leader
--------------
Project Leader:
Binlin Wu

Email: wub1@southernct.edu
Mobile Phone: 917-292-6794
Work Phone: 6465938848

Project Personnel
-----------------
Mentor(s):
Neil McGlohon (471), Jason Smith (584)

Student-facilitator(s):
Andrew Cheng (561)

Mentee(s):
{Empty}


Project Information
-------------------
Project Description:
In this project, we will attempt to use convolutional neural networks (CNNs) deep learning methods to analyze and classify fluorescence images collected from two different types of kidney tumor tissue slides. 

Project Information Subsection
------------------------------
Project Deliverables:
If successful, this project will develop a set of parameterized CNNs able to analyze a set of images and achieve high classification accuracy.

Project Deliverables:
{Empty}

Student Research Computing Facilitator Profile:
Upper level undergrad or grade, physics or computer science students with programming background.


Mentee Research Computing Profile:
{Empty}

Student Facilitator Programming Skill Level: Practical applications
Mentee Programming Skill Level: {Empty}
Project Institution: Southern Connecticut State University
Project Address:
501 Crescent Street, ASB/SCI 045
New Haven, Connecticut. 06515

Anchor Institution: CR-Yale
Preferred Start Date: 10/01/2020
Start as soon as possible.: No
Project Urgency: Already behind3Start date is flexible
Expected Project Duration (in months): {Empty}
Launch Presentation: {Empty}
Launch Presentation Date: 03/10/2021
Wrap Presentation: {Empty}
Wrap Presentation Date: 05/12/2021
Project Milestones:
- Milestone Title: Milestone 1
  Milestone Description: Adopt a model and make it run on a GPU
  Completion Date Goal: 2020-11-01
- Milestone Title: Milestone 2
  Milestone Description: Optimize the model for the image set
  Completion Date Goal: 2020-12-01
- Milestone Title: Milestone 3
  Milestone Description: Apply the model to classify images and produce results. Give brief wrap-up presentation.
  Completion Date Goal: 2021-02-01

Github Contributions: {Empty}
Planned Portal Contributions (if any):
{Empty}

Planned Publications (if any):
Publication is expected if good results are achieved.

What will the student learn?:
Cancer diagnosis, optical imaging, biomedical image analysis, deep learning, GPU usage

What will the mentee learn?:
{Empty}

What will the Cyberteam program learn from this project?:
{Empty}

HPC resources needed to complete this project?:
GPU-equipped nodes

Notes:
Interested in using AiMOS at Rensselaer



Final Report
------------
What is the impact on the development of the principal discipline(s) of the project?:
This CAREERS project has made a significant impact on my research medical image analysis. The implementation of parallel computing using an HPC system is crucial in the project. Since the program and code developed in the project can also be used in other similar projects that involve image analysis, the impact the project has made is not limited to this project, but on a much broader area. 

What is the impact on other disciplines?:
This research is highly interdisciplinary. The problem we are dealing with is essentially common in any area where researchers study a substance based on image analysis. The work done in this project will definitely make an impact on other disciplines. The development of such technologies of image analysis using deep learning can be used in biomedical areas such as cancer diagnosis, remote sensing such as radar or lidar technologies, biometric analysis, nanotechnology for material characterization, astronomical research, etc.

Is there an impact physical resources that form infrastructure?:
The researchers including myself who have participated in the program have learned a lot through this program, such as what resources are available in this region, experts we can seek help from, and the mechanism of the program. The successful experience from this program will inspire us to acquire physical resources in our own institution to support research. With collaboration, the available resources we have learned might help us to develop such infrastructure.

Is there an impact on the development of human resources for research computing?:
Through this research, the students have been trained to be research facilitators to help researchers in different areas to perform crucial modeling and computing tasks in their research. The students got trained not only in the computing techniques using HPC but also learned how to work with researchers to implement computations to solve a scientific problem. It has definitely made a significant impact on the development of human resources for research computing. 

Is there an impact on institutional resources that form infrastructure?:
The successful experience of the student and the researcher will be a great demonstration to my colleagues and the administration, which may inspire my institution to arrange resources to support the development of infrastructure.  

Is there an impact on information resources that form infrastructure?:
Yes, as I have mentioned above, the program provided great information resources to the students and the researchers, who can play important roles to form infrastructure.

Is there an impact on technology transfer?:
N/A

Is there an impact on society beyond science and technology?:
The research is highly interdisciplinary. It covers multiple areas such as cancers, optical imaging, image processing, statistical analysis, artificial intelligence etc. The technology being developed can be transferred to many other areas with important applications and make an impact on society beyond science and technology.

Lessons Learned:
N/A

Overall results:
I think the overall results of the project are outstanding. The program itself and the trained research facilitators are trying to fill the gap between the computing community and the general research community in which intensive computation plays a crucial role both in the short term and in the long term. Through the project, the researchers and the students all gained a lot from the program. The research (myself) can get the project accomplished with great resources as support including the computing experts and the physical resources. In the meantime, the information of the resources is another beneficial part to the researchers. The program provides an excellent learning opportunity to the student trainees with the support of mentors, physical resources, and networking opportunities, which can be an eye-opening or even life-changing opportunity. They have a chance to learn about computing techniques especially using HPC systems, performing computing tasks in scientific research, and even career paths that they have never thought about before. All these they have gained from the program can make a great impact on their future career.