visfeats.memorability_resmem

visfeats.memorability_resmem(image: Image) → float

Estimate the intrinsic memorability of an image.

The image is converted to RGB, preprocessed with the ResMem transformer (resized to 227 x 227 and normalized), and passed through the pretrained ResMem network, a residual neural network trained to predict human memorability scores [1]. The model is loaded once per process and reused across calls.

This feature requires the optional resmem package, which is not installed by default (pip install resmem).

Parameters:

imageImage

Input image in any Pillow mode; converted to RGB internally.

Returns:

float

Predicted memorability score in [0, 1]; higher values indicate a more memorable image.

Raises:

ImportError

If resmem package is not installed.

References

[1]

Needell, C. D., & Bainbridge, W. A. (2022). Embracing new techniques in deep learning for estimating image memorability. Computational Brain & Behavior. https://doi.org/10.1007/s42113-022-00126-5