menpofit is Menpo Project's Python package for building, fitting and manipulating state-of-the-art 2D deformable models.

We strongly advise you to first visit the Basics section in order to understand the fundamental concepts and assumptions that are made in menpofit, before reading about the actual methods.

menpofit provides solutions to the following problems:

Affine Image Alignment

An Affine Image Alignment algorithm aims to find the optimal alignment between an input image and a template image with respect to the parameters of an affine transform.

The methods that are implemented in menpofit are:

• Optimization Algorithms: Forward Additive, Forward/Inverse Compositional

Deformable Image Alignment

Deformable Image Alignment aims to get the optimal alignment between an input image and a template image with respect to the parameters of a statistical parametric shape model.

The methods that are implemented in menpofit are:

• Active Template Model (ATM)

Landmark Localization

A deformable object is commonly represented with a set of landmarks which correspond to semantically meaningful parts.

Landmark Localization is the problem of localizing the landmark points that correspond to a deformable model in an input image.

The models that are implemented in menpofit are:

Finally, please see the References for an indicative list of papers that are implemented in menpofit.

Deformable Objects

Note that all the examples of this menpofit user guide are based on the human face for demonstration purposes. The Menpo Project is not specific to the human face and can be used for any kind of deformable object, such as the human body (skeleton), human hand, cat face and car sideview.