Survival Guide from Matlab to Python
About This Guide
This is a guide for switching from Matlab to Python, mainly for researchers in deep learning and computer vision, especially those working with biomedical images. The document is adapted from my personal study note and is organized by different application scenarios. (Important Note: Most of the solutions provided in this guide is only one possible realisation.)
Background
What I used before, as a CS Ph.D. student:
- Matlab (for implementing biomedical image analysis algorithm and preparing data for deep learning experiments)
- Torch (for deep learning research)
Now, I work mostly in Python with various packages:
- SimpleITk for classic biomedical image analyis routines (io, basic processing, quantitative analysis, etc.)
- pandas for non-image data io.
- PyTorch for deep learning research
Task-by-task Guide
1. Read image data
The very first step of all image processing tasks is to read images.
2. Read non-image data
Sometimes, you may want to read data from csv files, e.g., batch processing and the data directories are saved in csv, or reading classification labels.
3. Playing with data directory
4. Python 101: Loop, if/else, string, etc.
5.
to be continued …