PolyChar

Official codebase for "Computer Vision for Polymer Characterisation using Lasers"

PolyChar is a simple laser-based platform that combines computer vision and deep learning models to classify the solubility of different polymeric compounds across a range of solvents. Using the results obtained from the solubility screening method, Hansen Solubility Parameters (HSP) of the polymers using are determined using an optimisation algorithm. Additionally, a Convolutional Neural Network regression model is also used to estimate the size of polymeric nanoparticles between 20-470 nm.

Method

(Designed to be compatible with light-colored backgrounds.)

Installation

To install the repository, follow these steps (Note that this will install the CPU version of torch only):

1. Clone the repository:

   git clone https://github.com/sduynk/Polymer_characterisation.git
   
   cd Polymer_characterisation

2. Create a conda environment and ensure it is activated:

   conda create --name polychar_env python=3.9
   conda activate polychar_env

3. Install the required dependencies:

   pip install -r requirements.txt

4. Verify the installation:

   python -m unittest discover

If you want to install torch with CUDA for GPU acceleration, afterwards you can do the following.

5. Uninstall torch and torchvision from the environment

   pip uninstall torch torchvision
   

6. Install torch and torchvision with CUDA support e.g. cu121 (see notes for CUDA version):

   pip install torch==2.3.0+cu121 torchvision==0.18.0+cu121 --extra-index-url https://download.pytorch.org/whl/cu121

Note:

• The CUDA version (e.g. cu121) should match your GPU and driver.
• For most users, you do not need to install the CUDA toolkit separately.
• If you have a different GPU or CUDA version, see PyTorch's official installation guide.
• While efforts have been made towards bit-accurate reproducibility, differences in drivers and cuda toolkits, among other factors, will likely yield slightly different results.

Datasets

The necessary datasets to run the code are hosted at https://zenodo.org/records/15480040. Replace the empty folders for...

• Particle Size Regression/Particle_Size-data
• Solubility Classification/Solubility-data
• Hansen Solubility Parameters/Hansen_Solubility_Parameters-data

With the corresponding folders in Zenodo and you should be good to go!

Solubility Results

Running solubility.ipynb on the solubility dataset should give the following (or similar) results (ommitting @3 and @2 metrics).

Model	F1 Score@4	Accuracy@4	Precision@4	Recall@4
ResNet18	0.868±0.045	0.901±0.037	0.865±0.054	0.886±0.047
EfficientNet B0	0.853±0.055	0.891±0.043	0.850±0.063	0.876±0.061
ConvNext Tiny	0.865±0.045	0.904±0.035	0.869±0.051	0.878±0.050

Hansen Solubility Parameters Results

To obtain HSP Results, cd into the Hansen Solubility Parameters directory and run python Genetic_algorithm.py

Polymer	Conc. (% w/v)	δD (GT)	δP (GT)	δH (GT)	δD (Pred)	δP (Pred)	δH (Pred)	R₀	ED	PED (%)
PMMA	5	18.6	10.5	5.1	17.4	10.4	3.1	9.2	2.4	11
PS	5	18.5	4.5	2.9	18.1	3.9	5.7	4.5	2.9	15
PVP	5	17.5	8.0	15.0	20.0	12.6	14.1	13.4	5.3	22
PCL	5	17.7	5.0	8.4	18.3	10.5	5.0	9.6	6.5	32

Particle Size Results

• Running particle_size.ipynb should give the following (or similar) results for PPSNet - MLP (ReLU).
• Running polynomial_regression.ipynb can be used to obtain the polynomial_regression results.

Method	MAE (nm)(mean ± std)	RMSE (nm)(mean ± std)	R²(mean ± std)
PPSNet - MLP (ReLU)	9.53 ± 4.27	15.60 ± 7.58	0.99 ± 0.01
PPSNet (no conditioning)	22.25 ± 3.97	32.01 ± 6.95	0.93 ± 0.04
Polynomial Regression	32.55 ± 6.67	47.81 ± 9.58	0.87 ± 0.03
EfficientNet - MLP (Sine)	11.60 ± 3.07	20.13 ± 5.95	0.98 ± 0.01

Project Structure

The project is organized as follows:

PolyChar/
├── Solubility Classification/
│   ├── Solubility-Data/        
│   │   └── annotation.csv      # Data found in Zenodo
│   ├── results/                # ^
│   ├── Trained_Models/         # ^
│   ├── dataloaders.py 
│   ├── models.py 
│   ├── solubility.ipynb
│   ├── summarize_results.ipynb
│   ├── train.py 
│   └── utils.py 
├── ParticleSize Regression/
│   ├── Particle_Size-Data/
│   │   └── annotation.csv      # Data found in Zenodo
│   ├── results/                # ^
│   ├── Trained_Models/         # ^
│   ├── dataloaders.py 
│   ├── figures.ipynb
│   ├── particle_size.ipynb
│   ├── polynomial_regression.ipynb
│   ├── ps_models.py
│   ├── train_regression.py 
│   └── utils.py 
├── Hansen Solubility Parameters/
│   ├── Genetic_algorithm.py
├── README.md
└── requirements.txt

Other Files

• README.md: This file, providing an overview of the project.
• requirements.txt: Lists the dependencies required to run the project.

Authors

George Killick (george.killick@liverpool.ac.uk) Seda Uyanik (seda.uyanik@liverpool.ac.uk)

License

Distributed under the Unlicense License. See LICENSE.txt for more information.