Generalized assorted pixel camera: postcapture control of resolution, dynamic range, and spectrum

Area of Science:

• Computer Vision
• Image Processing
• Computational Photography

Background:

• Traditional cameras capture fixed image properties.
• Color filter arrays in cameras can lead to under-sampling and aliasing artifacts.
• Reconstructing high-quality images from under-sampled data is challenging.

Purpose of the Study:

• To introduce the Generalized Assorted Pixel (GAP) camera concept.
• To develop methods for optimizing color filter array design and image reconstruction.
• To demonstrate flexible control over image properties (resolution, dynamic range, spectral detail) from a single capture.

Main Methods:

• Comprehensive optimization of the color filter array's spatial and spectral layout.
• Novel algorithm for reconstructing under-sampled image channels, minimizing aliasing.
• Demonstration of generating various image types (monochrome, HDR, RGB, multispectral) from a single GAP image capture.

Main Results:

• Successful optimization of GAP camera color filter array design.
• Effective reconstruction of under-sampled channels with minimized aliasing artifacts.
• Demonstrated user control over the tradeoff between spatial resolution, dynamic range, and spectral detail.
• Validation of GAP camera performance through extensive simulations using real-world multispectral scenes.

Conclusions:

• The GAP camera offers unprecedented post-capture flexibility in image property control.
• The developed reconstruction algorithm effectively mitigates aliasing from under-sampled filter arrays.
• The GAP camera system provides a versatile platform for generating diverse image types from a single capture.