Introduction

pyMLV is the Python version of the Mid-level Vision Toolbox (MLVToolbox), created by the Aravind Narayanan in BWLab at the University of Toronto. The toolbox facilitates the extraction and analysis of structural properties in images such as contours' orientation, length, curvature, and junctions. It supports a variety of research tasks including the identification of perceptual organization cues like mirror symmetry, ribbon symmetry, and taper symmetry in naturalistic images.

Usage

To get started, clone the repository and navigate to the demos folder. You can execute the Jupyter notebooks within to see examples of the toolbox's capabilities in action:

• Demo_ContourFeatures: Demonstrates extraction of contour features.
• Demo_MedialAxis: Shows how to compute and visualize the medial axis of images. Note that the medial axis demo requires additional components due to MATLAB dependencies.

Updates and Development

The toolbox is actively maintained, and updates are periodically released. Most functions required for the contour feature and medial axis demonstrations are included, and further enhancements are continually being developed.

References

If you utilize this toolbox in your research, please consider citing the following papers:

• Walther, D. B., Farzanfar, D., Han, S., & Rezanejad, M. (2023). The mid-level vision toolbox for computing structural properties of real-world images. Frontiers in Computer Science, 5. doi: 10.3389/fcomp.2023.1140723
• Rezanejad, M., Downs, G., Wilder, J., Walther, D. B., Jepson, A., Dickinson, S., & Siddiqi, K. (2019). Scene categorization from contours: Medial axis based salience measures. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 4116-4124).
• Walther, D. B., & Shen, D. (2014). Nonaccidental properties underlie human categorization of complex natural scenes. Psychological science, 25(4), 851-860.
• Rezanejad, M., & Siddiqi, K. (2013). Flux graphs for 2D shape analysis (pp. 41-54). Springer London.