altmain.ProjectionSplit Class Reference

Handles splitting images for dual-projector setups with overlap blending. More...

Public Member Functions
	__init__ (self)
	Initializes the ProjectionSplit processor with NumPy backend.
	process_images (self, int overlap=75, str blend_type="linear")
	Main public interface for processing static images.

Public Attributes
	cfg
	Configuration manager instance.
	image_left
	Processed left projector output image.
	image_right
	Processed right projector output image.
	image_main
	Original input image buffer for reference.

Protected Member Functions
Tuple[np.ndarray, np.ndarray]	_generate_alpha_curves (self, int overlap, str blend_type)
	Generates alpha blending curves for overlap region.
None	_apply_blending (self, np.ndarray image, int overlap, str blend_type)
	Applies alpha blending to split images using NumPy operations.

Detailed Description

Handles splitting images for dual-projector setups with overlap blending.

This class is responsible for:

• Loading and processing input images for dual-projector setups
• Generating mathematical alpha blending curves for smooth transitions
• Splitting images into left/right sections with configurable overlap
• Applying blending using pure NumPy operations (no GPU dependency)
• Saving processed outputs to disk for projector display

Constructor & Destructor Documentation

__init__()

altmain.ProjectionSplit.__init__

(

self

)

Initializes the ProjectionSplit processor with NumPy backend.

Constructor that sets up the processor by:

• Initializing image storage buffers as None
• Loading configuration from config.ini file
• Using pure NumPy operations for cross-platform compatibility

Postcondition

self.image_left, self.image_right, self.image_main are initialized to None

self.cfg contains loaded configuration from config.ini

Exceptions

FileNotFoundError

if config.ini cannot be found

Member Function Documentation

_apply_blending()

None altmain.ProjectionSplit._apply_blending (   self, np.ndarray  image, int  overlap, str  blend_type  )

protected

Applies alpha blending to split images using NumPy operations.

Core processing method that:

1. Converts 3-channel images to 4-channel (BGRA) if needed
2. Splits image into left and right sections with specified overlap
3. Generates alpha blending curves
4. Applies curves to alpha channels of overlap regions
5. Stores results in instance buffers and saves to disk

Parameters

image	Input image as numpy array (BGR or BGRA format)
overlap	Width of overlap region in pixels (must be even for symmetric split)
blend_type	Type of blending curve to apply

Postcondition

self.image_left and self.image_right contain processed images

left.png and right.png are saved to working directory

_generate_alpha_curves()

Tuple[np.ndarray, np.ndarray] altmain.ProjectionSplit._generate_alpha_curves (   self, int  overlap, str   blend_type  )

protected

Generates alpha blending curves for overlap region.

Creates mathematical curves that determine transparency values across the overlap zone using three supported blending types:

• Linear: Straight-line fade with configurable slope parameter
• Quadratic: Parabolic fade (x²) for smoother transition
• Gaussian: Bell curve fade using error function for natural blending

Parameters

overlap	Width of the overlap region in pixels
blend_type	Type of blending curve: "linear", "quadratic", or "gaussian"

Returns

Tuple of (left_curve, right_curve) where each is a 1D numpy array of alpha values (float32) ranging from 0.0 to 1.0 across overlap width

Exceptions

ValueError

if an unsupported blend_type is provided

process_images()

altmain.ProjectionSplit.process_images	(		self,
		int	overlap = 75,
		str	blend_type = "linear"
	)

Main public interface for processing static images.

Loads input image from disk and applies splitting and blending operations. This is the primary method called by the GUI controller for image processing.

Parameters

overlap	Overlap width in pixels (default: 75)
blend_type	Blending curve type (default: "linear")

Exceptions

FileNotFoundError	if input.png cannot be found in working directory
ValueError	if blend_type is not supported

Postcondition

self.image_main contains the loaded input image

self.image_left and self.image_right contain processed outputs

left.png and right.png are saved to working directory

---

The documentation for this class was generated from the following file:

• Doxygen_Final/altmain.py