Project Info
Here is some information about the final project:
- You can work alone or in a group (up to 4 people), required effort will scale with # of people
- Select a “base” dataset (see examples below)
- You’ll need to simulate parameters of a physical (imaging) system with a base dataset to form a modified dataset
- You’ll then need to train a deep neural network, with the modified dataset, and the physical parameters in mind
- Please see slides from class for a number of examples and discussions
Here is a link to the final project instructions: Final project instructions
Here is a link to a template for the final project research-style paper: Template for research-style paper
Here is a link to a template for the final project short webpage: Template for final project webpage
Here is a link to the permission form: Permission form
Using the above files, you’ll thus turn in the following:
- An 7-minute presentation with slides that you will give to the class during the final period
- A short research-style paper (4 - 6 pages on average) that includes an abstract, introduction, related work, methods, results, a discussion, references and at least 3 figures.
- The project’s source code
- A completed web template containing the main results from the research paper
- A completed permission form
Here are some important dates:
- Final code and presentation due via email: Thursday May 2, 5pm
- Final presentation time slots: Thursday May 2, 7pm - 10pm
- Final presentation paper write-up, website template and permission form due: Saturday May 4 at 11:59pm
- Final presentations will occur via Zoom, please coordinate with team members and practice ahead of time!
Some datasets that might be helpful:
- Grand Challenge datasets - lots of datasets to choose from here!
- Label free prediction datasets from Allen Institute
- HAM10000 pigmented skin lesion dataset
- In-silico labeling dataset
- Publically labeled Cell Atlas dataset
- Large number of datasets from the Broad Institute
- Biological image synthesis
- IDR Open Microscopy datasets
- Tensorflow Datasets
- Fourier MNIST and malaria datasets
- Scene Flow datasets
- NYU depth dataset
- Scattering media dataset from Rice
- Flatcam Face dataset
- DeepLesion CT dataset
- The Cancer Imaging Archive
- Kaggle
- Augmented Cognition Laboratory(ACLAB) - northeastern university
- Zenodo
- Haravard Dataverse
- Mendeley Datasets
- ISIC challenge datasets
- Broad Institute
- Medical Imagery Datasets
- MICCAI
- Deep Multi-modal Object Detection and Semantic Segmentation for Autonomous Driving
- Multispectral Image Recognition
- COVID-NET
- Awesome Data
- Refocusing and Super Resolution dataset for Cytology, Original Paper Link
Here are a few useful datasets that were used in previous classes:
- Malaria-infected blood cells under variable-angle illumination
- Annotated microscope images of malaria-infected blood, tuberculosis in sputum, and other parasites
- Broad Institute Annotated Cell Image Datset
- Microwave Detection of Hand Gestures from the Smith Lab at Duke
- High Dynamic Range Color Image Dataset
- Optical Coherence Tomography Annotated Datasets from Farsiu Lab at Duke
- Chest X-ray images (pneumonia)