Stmicroelectronics Asia Pacific Pte Limited

Stmicroelectronics S.a.

Stmicroelectronics (shenzhen) R&d Co. Ltd.

Group Des Ecoles Des Telecommunications

Lucas Hui

Yong Huang

Xiaoyun Deng

Haiyun Wang

Fritz Lebowsky

Patricia Wei Yin Chiang

Jean-Luc Danger

Patricia Chiang

Jean-Michel Bard

Guillaume Militello

Anna Raffalli

Christophe Cunat

Various embodiments provide tone mapping of images and video from one dynamic range to an available dynamic range of a display device while preserving or enhancing image details. According to one embodiment, an enhanced tone mapping module is configured to decrease a luminance of an image or video from a high dynamic range to a standard dynamic range. Conversely, according to one embodiment, an enhanced inverse tone mapper to increase a luminance of an image or video from a standard dynamic range to a high dynamic range.

Enhanced tone mapper for high dynamic range images and video

Ghosting of images can be produced by the use anaglyph techniques to produce stereoscopic images. Ghost effects can be produced when filtering is applied to images by displays such as televisions or monitors. Techniques and apparatus are described for reducing or cancelling ghost effects. Anaglyph ghost reduction information can be produced to reduce or cancel ghost effects.

Anaglyph ghost cancellation

A video de-interlacing device includes a video data input interface to receive a stream of interlaced video data, a motion detector coupled to the video data input interface to produce motion detection video data, and a video data output interface to pass de-interlaced video data generated from the motion detection video data. The video de-interlacing device also includes a recursive motion processor to detect a still motion condition in at least one area of the motion detection video data. The recursive motion processor includes a local still detector, a large area motion detector, and a feathering detector. The recursive motion processor is arranged to produce motion information used to crop the motion detection video data based on the detected still motion condition.

Recursive motion video de-interlacing device

A method, system and apparatus for adaptive pre-filtering includes: extracting tuning parameters from video data in a video encoding process; processing the tuning parameters to generate control parameters representative of properties of the video data; coupling the control parameters to a filter response; generating filter coefficients according to the filter response; and filtering the video data using the filter coefficients prior to video encoding. The filtering of video data may thus be adapted to one or more tuning parameters of the video data and the encoding process. The filtering may decrease the complexity of encoding by an attenuation of high frequency signals, effectively decreasing the quantization step and reducing compression artifacts.

Adaptive pre-filtering system

Digital noise removal methods and systems for picture quality improvement on next generation high quality set-top-box and digital TV SOCs is disclosed with particular to a de-blocking method and system that is effective on on-grid and off-grid blocky artifacts generated by lossy video compression standards. Blocky artifacts are filtered by extracting motions vectors for a current pixel block, calculating off-grid positions based on the extracted motion vectors, and applying de-blocking filtering along the calculated off-grid positions. The methods and systems can further skip filtering on-grid blocky artifacts to be filtered by an on-grid de-blocking filter known in the art.

Motion compensated de-blocking

A system adjusts contrast of an anaglyph image and includes a contrast adjuster processor that generates a contrast-adjusted first luma signal and second luma signal and a contrast-adjusted third luma signal from a first input image signal and second input image signal. An upstream summing node determines a first difference between the first luma signal and third luma signal and a downstream summing node subtracts the first difference from the first signal component of the first input signal. A fine adjuster circuit receives a signal for the first difference and applies a fine adjustment as a gain adjustment, an offset adjustment, and a horizontal and vertical shift.

Contrast adjusted anaglyph for reducing ghosting artifacts

A method and apparatus are provided for detecting banding noise in a digital signal representative of an image. The method includes determining, by a banding noise detector, a count of increment steps in pixel values and a count of decrement steps in pixel values along a filter direction in a neighborhood of a current pixel of the image, checking, by the banding noise detector, if the count of increment steps or the count of decrement steps in the neighborhood of a current pixel exceeds a step threshold value, and classifying, by the banding noise detector, the current pixel as being located in the banding noise zone if the count of increment steps or the count of decrement steps does not exceed the step threshold value.

Banding noise detector for digital images

A system for determining an aspect ratio of image content based on an analysis of the content. In an embodiment, an analyzer is configured to receive a data input corresponding to an image in a stream of images that constitute a video sequence of images. The analyzer is further configured to determine a mathematical representation of the image content based on a power spectrum analysis of vertical components of the image in comparison to a power spectrum analysis of horizontal components of the image. Based on this comparison of the vertical frequency components to the horizontal frequency components, a determination about the original aspect ratio of the image may be determined. This determination may be used by a video processor to correctly apply aspect ratio conversion for final image output.

Content-based aspect ratio detection

A method and apparatus are provided for filtering banding noise in a signal representative of an image. The method includes detecting, by a banding noise detector, banding noise in a neighborhood of a current pixel of the image, determining, by an adaptive filter weight decision unit, a number of banding steps in the neighborhood of the current pixel, determining, by the adaptive filter weight decision unit, a difference between a current pixel value and a previous output value, selecting, by the adaptive filter weight decision unit, a filter weight based on the number of banding steps, the difference between the current pixel value and the previous output value, and the detected banding noise, and filtering, by a recursive filter, the current pixel value according to the selected filter weight.

Recursive de-banding filter for digital images

The present invention relates to systems and methods for inverse telecine or video de-interlacing for picture quality improvement on set-top-box and TV products. The system comprises a film mode detector at the picture or sequence level, a global mixed video and film content detector at the region, picture, or sequence level on top of the detected film content, and a local video content detector at pixel level on top of the detected mixed video and film content. Inverse telecine processing is applied on detected film content fading in with a locally de-interlaced local video content. The invention further provides an apparatus and method for globally detecting mixed video and film content at region, picture, or sequence level. Such apparatus and method comprise a plurality of detectors for robustness and increased detection accuracy.

Apparatus and method for inverse telecine with local video de-interlacing

Systems and methods are disclosed for determining the perceptibility of noise in a block of images and/or video. The systems and methods may compute a mask value for the block using a block masking generator. The mask value may indicate the perceptibility of noise in the block. The mask value may be computed using a normalized activity value and/or a texture value for the block. The normalized activity value may indicate the relative activity in the block as compared to the activity in the image and/or video. The texture value may indicate the strength and/or number of edges in the block.

Perceptual block masking estimation system

The appearance of image details can be preserved and/or enhanced by applying contrast adaptive gain to the high spatial frequency component of the luminance information. The image details in bright and/or dark regions can be further boosted by applying a local mean adaptive gain. The contrast transfer mapping curve for luminance contrast enhancement can be re-scaled to account for the applied gain. The re-scaling may be performed from frame to frame of displayed video. The re-scaling may be temporally controlled for subsequent frames to make the re-scaling change gradually to prevent flickering.

Image details preservation and enhancement

A method and apparatus are disclosed for identifying and removing banding artifacts (i.e., false contours) resulting from insufficient bit depth caused by digital image quantization, conversion, and/or compression. This method includes: explicitly identifying texture block and flat block; de-termination of filter window sizes with the consideration of handling transitions between texture block and flat block; de-banding filtering with edge protection; and noise shaping according to de-banding filter result.

General banding and codec banding artifact removal

The methods and systems of this invention allow for independent adaptive control of ringing and overshoot effects in 2-dimensional array interpolation processes, including in image and video rescaling and analysis. The methods and systems can use either a column-wise or a row-wise interpolation, or a combination thereof. Each uses a respective preliminary interpolation of data, followed by ringing and/or overshoot control. Controllable parameters allow variability in the amount of ringing and/or overshoot retained in the interpolated data. The ringing and overshoot controls apply a local analysis of the data to adjust the preliminary interpolation results. The methods may be repeated iteratively, for example, to obtain a desired rescaling of an image data array.

2D ringing and overshoot control in image rescaling

Provided are a digital video rescaling system, a method of rescaling video images, and a chip comprising a computer executable medium embedded therein computer executable instructions for rescaling video images.

Growing community of inventors

Singapore, Singapore

Lucas Hui