Abstract: Systems and methods are provided for merging multiple images to produce a single fused image having desirable image characteristics derived from the multiple images. For example, the system may determine image characteristics for first and second images. The image characteristics may be related to contrast, exposure, color saturation, and the like. Based on the image characteristics, the system may generate a combined luma weight map. The system may decompose the first and second images and the combined luma weight map. In an example, the first image, the second image, and the combined luma weight map may be represented as scale-space representations having multiple scales or levels. The system may merge the decomposed representations of the first and second images and the combined luma weight map to form a decomposed representation of the fused image. The system may generate the actual fused image from the decomposed representation of the fused image.

Abstract: Systems and methods are provided for merging multiple images to produce a single fused image having desirable image characteristics derived from the multiple images. For example, the system may determine image characteristics for first and second images. The image characteristics may be related to contrast, exposure, color saturation, and the like. Based on the image characteristics, the system may generate a combined luma weight map. The system may decompose the first and second images and the combined luma weight map. In an example, the first image, the second image, and the combined luma weight map may be represented as scale-space representations having multiple scales or levels. The system may merge the decomposed representations of the first and second images and the combined luma weight map to form a decomposed representation of the fused image. The system may generate the actual fused image from the decomposed representation of the fused image.

Abstract: A method for making a slice cap decision during encoding of a video sequence in a video encoding pipe is provided that includes determining a prediction mode for a macroblock, computing an estimated encoded size of the macroblock based on an adaptive macroblock size estimation model, and making a slice cap decision for a current slice based on the estimated encoded macroblock size, wherein the slice cap decision is made prior to entropy coding the macroblock.

Abstract: A method for making a slice cap decision during encoding of a video sequence in a video encoding pipe is provided that includes determining a prediction mode for a macroblock, computing an estimated encoded size of the macroblock based on an adaptive macroblock size estimation model, and making a slice cap decision for a current slice based on the estimated encoded macroblock size, wherein the slice cap decision is made prior to entropy coding the macroblock.