Skip to content

supersonicMaclaurin/AI4Science_exp

Repository files navigation

AI4Science

ZS-DeconvNet: Image Super-Resolution in AI4Science

ZS-DeconvNet is an zero-shot deconvolution network crafted to elevate the resolution and signal-to-noise ratio (SNR) of microscopy images, thereby significantly improving the clarity and quality of biological observations.

Experiment setup

  • PC with an Intel Core i7-10700 processor and an NVIDIA RTX 3090ti GPU
  • TensorFlow 2.5.0 and
  • Python 3.9.7

Training was conducted using the Adam optimizer with an initial learning rate of 5e-4, which decays by a factor of 0.5 every 10,000 iterations. A batch size of 4 was utilized for the training process, which generally encompasses 50,000 iterations.

file structure

|--- reproduce_model // my reproduced model

|--- test_results // my test results of wide_field_images and Structured Illumination Microscopy (SIM) imaging of clathrin-coated pits (CCPs)

About

深度神经原理课程实验-AI4Science

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published