Autofocusing via Combined Horizontal and Vertical Difference Images

The patent combines phase-detection autofocus and contrast autofocus into a unified system. The phase-detection module provides initial focusing speed, while the contrast module refines accuracy in challenging conditions. The system dynamically switches between or combines both methods based on scene characteristics, thereby achieving both fast focusing speed and high accuracy across diverse lighting and texture conditions.

The patent introduces an image processing module as an intermediary that analyzes scene characteristics (illumination, contrast, texture) and determines the appropriate focusing strategy. This mediator processes the raw image data and controls the switching between phase-detection and contrast autofocus modes, enabling the system to adapt to varying conditions and maintain both speed and accuracy.

The patent makes the image sensor multi-functional by enabling it to perform both image capture and phase-detection autofocus. By utilizing the existing pixel array for dual purposes through computational methods (comparing pixel responses to determine focus), the system eliminates the need for a separate dedicated phase-detection sensor, thereby reducing device complexity and cost while maintaining autofocus capability.

The patent replaces the mechanical/optical dedicated phase-detection sensor with a computational approach using the standard image sensor. Instead of relying on specialized hardware, the system uses software-based phase-detection algorithms that process images from the regular pixel array, substituting a complex mechanical system with a simpler computational solution.