tSPM-Probabilistic-Scanpaths
Repository for the work "A Probabilistic Time-Evolving Approach to Scanpath Prediction"
Requeriments
You can install an environment with all required dependencies using the environment.yml file in Anaconda.
Inference
The current version of the repository includes a basic, yet functional version to generate scanpaths from a single image using our model.
Usage
There is currently one mode of usage for this code:
python main.py --mode inference
Inference parameters
image_path: The path from the image from which to generate scanpaths.
path_to_save: The path to save the generated scanpaths.
n: The number of scanpaths to generate (default = 20).
th: The probabilistic threshold (default = 0.5, see the paper for details).
This will read an image from image_path = "data/test.jpg" and generate a set ofn scanpaths that will be saved in path_to_save = "test/".
Training the model
This option will be available soon.