Abstract: The disclosure includes a system and method for fake signature detection including identifying, using one or more processors, within a document image a signature portion, the signature portion representing a signature; applying, using the one or more processors, one or more fake signature detection models to the signature portion; determining, using the one or more processors and the one or more fake signature detection models, whether the signature represented in the signature portion is a fake signature; and presenting, using the one or more processors, the determination.

Abstract: Input minimal noise levels of are computed over input codeword bins based on image content in input images of an input bit depth. The minimal noise levels are adjusted to generate approximated minimal noise levels of a higher bit depth. The approximated minimal noise levels are used to generate per-bin bit depths over the input codeword bins. The input codeword bins are classified into first codeword bins that have relatively high risks of banding artifacts and some other input codeword bins that have relatively low or zero risks of banding artifacts based on the per-bin bit depths. Portions of bit depths from the other input codeword bins are moved to the first input codeword bins to generate modified per-bin bit depths. A forward reshaping function constructed from the modified per-bin bit depths is used to reshape the input images into reshaped images used to generate output images.

Abstract: In a method to improve the dynamic range of high-dynamic range (HDR) signals using an enhancement layer, a piecewise-linear inter-layer predictor and a residual masking operator are applied. The generation of the piecewise-linear inter-layer prediction function is based on a computed scene-significance histogram based on the average of frame-significance histograms indicating pixel values where coding artifacts are most likely to occur. For each segment in the prediction function, its slope is inversely proportional to a measure of energy in the segment under the scene-significance histogram. Bit rate constrains for the enhancement layer are also taken into consideration in determining the piecewise-linear prediction function.

Abstract: Input minimal noise levels of are computed over input codeword bins based on image content in input images of an input bit depth. The minimal noise levels are adjusted to generate approximated minimal noise levels of a higher bit depth. The approximated minimal noise levels are used to generate per-bin bit depths over the input codeword bins. The input codeword bins are classified into first codeword bins that have relatively high risks of banding artifacts and some other input codeword bins that have relatively low or zero risks of banding artifacts based on the per-bin bit depths. Portions of bit depths from the other input codeword bins are moved to the first input codeword bins to generate modified per-bin bit depths. A forward reshaping function constructed from the modified per-bin bit depths is used to reshape the input images into reshaped images used to generate output images.

Abstract: In a method to improve the dynamic range of high-dynamic range (HDR) signals using an enhancement layer, a piecewise-linear inter-layer predictor and a residual masking operator are applied. The generation of the piecewise-linear inter-layer prediction function is based on a computed scene-significance histogram based on the average of frame-significance histograms indicating pixel values where coding artifacts are most likely to occur. For each segment in the prediction function, its slope is inversely proportional to a measure of energy in the segment under the scene-significance histogram. Bit rate constrains for the enhancement layer are also taken into consideration in determining the piecewise-linear prediction function.

Abstract: Statistical values are computed based on received source images. An adaptive reshaping function is selected for one or more source images based on the one or more statistical values. A portion of source video content is adaptively reshaped, based on the selected adaptive reshaping function to generate a portion of reshaped video content. The portion of source video content is represented by the one or more source images. An approximation of an inverse of the selected adaptive reshaping function is generated. The reshaped video content and a set of adaptive reshaping parameters defining the approximation of the inverse of the selected adaptive reshaping function are encoded into a reshaped video signal. The reshaped video signal may be processed by a downstream recipient device to generate a version of reconstructed source images, for example, for rendering with a display device.

Abstract: Statistical values are computed based on received source images. An adaptive reshaping function is selected for one or more source images based on the one or more statistical values. A portion of source video content is adaptively reshaped, based on the selected adaptive reshaping function to generate a portion of reshaped video content. The portion of source video content is represented by the one or more source images. An approximation of an inverse of the selected adaptive reshaping function is generated. The reshaped video content and a set of adaptive reshaping parameters defining the approximation of the inverse of the selected adaptive reshaping function are encoded into a reshaped video signal. The reshaped video signal may be processed by a downstream recipient device to generate a version of reconstructed source images, for example, for rendering with a display device.