Texas Instruments Corporation

Yuji Itoh

Ngai-Man Cheung

Ho-Cheon Wey

Emi Arai

A wide dynamic range image sensor method combines the response of high-gain sensing cells and low-gain sensing cells with better linearity than the prior art. A search is made in successive central regions within the response curve of the high-gain and low-gain cells to find a highest slope linear fit. This highest slope and the corresponding offset are used in mixing the high-gain and low-gain responses to achieve a wide dynamic range.

Method for mixing high-gain and low-gain signal for wide dynamic range image sensor

This invention is a method for speeding up block matching based motion estimation for video encoder. The invention 1) calculates statistics for a candidate motion vector for a predetermined fraction of the pixels of a macroblock, 2) makes an early decision based on this preliminary cost function, and 3) terminates the block matching process without calculating the cost function for other pixels if the preliminary cost function is not less than a predetermined threshold. This early decision for goodness estimation provides an economy of processing load when a large portion of data is left untouched (i.e. unprocessed). The present invention employs feedback control to reduce the predetermined threshold for quick convergence upon each detection of a better match.

Motion estimation using early decision for goodness of estimation with quick convergence feedback

A method of variable length coding classifies each received symbol into one of a plurality of classifications having a corresponding variable length code table selected based upon a probability distribution of received symbols within the classification. The variable length codeword output corresponds to the received symbol according to the variable length code table corresponding to the classification of that received symbol. The plurality of classifications and the corresponding variable length code tables may be predetermined and fixed. Alternatively, the variable length code table may be dynamically determined with data transmitted from encoder to decoder specifying the variable length code tables and their configurations. Universal variable length code (UVLC) is used to code the symbols. Universal variable length code can instantiate to different variable length code tables with different parameters.

Entropy coding scheme for video coding

A quantizer employs a scaled integral inverse ratio division for quantization of an input T by a quantization step Q. The quantizer forms an integral approximation q of 2/Q by either trunc(2/Q) or round(2/Q). A multiplier multiplies the absolute value of T by the q. An adjustment factor is added alternatively to the absolute value of T prior to multiplication or to the product after multiplication. This adjustment factor minimizes errors near transition points in the quantization. This invention is applicable to both trunc(T/Q) and round(T/Q).

Fixed-point quantizer for video coding

This invention corrects chrominance misalignment that occurs during chrominance down-sampling and up-sampling. The invention extracts a binary index from the corresponding luminance signal. The binary index enables generation of a filter window. On down-sampling the filter window is applied to a block of source chrominance pixels which are filtered or not based upon the binary index. On up-sampling the binary index of the filter window for the target chrominance pixels determines which are filtered or not.

Joint pre-/post-processing approach for chrominance mis-alignment

Motion compensation as for video compression with motion estimation with either frame mode or field mode by initial prediction of motion mode and when prediction error is small enough truncate motion estimation search.

Video encoding

The objective of the invention is to make high-efficiency compressed coding for images. In coding of video signals in this application example, the edge is extracted in a frame of the input image in block . In block , classification is made into blocks where edge is present and blocks free of edge. In block , for the block where edge is present, based on the correlation between the edge direction and the distribution of the DCT coefficients, the optimum scanning order is determined and coding is performed by the scanning order.

Coding method for video signal based on the correlation between the edge direction and the distribution of the DCT coefficients

This invention is a method to automatically configure the universal variable length coding (UVLC). The method is applicable to code the syntax elements with a lot of symbols like the transform coefficients and motion vectors. The configuration problem includes a configuration method based on a genetic algorithm (GA). The method can be applied to on-the-fly configuration of codewords during video encoding, or to off-line training of code tables. This invention also includes several techniques to reduce the required operations for applications which have only limited processing power.

Automatic configuration of universal variable length code for video coding

A device and method for filtering out the noise generated due to coding of image data signals. The device has a threshold determining unit, a binary index unit, a filter selecting unit, and an adaptive filtering unit. The threshold determining unit 4 divides each pixel of the input image data into two gray levels. The binary indexes defined by the gray level are checked by a window with a prescribed size. If the region in the window is determined to be a homogeneous region, a heterogeneous region, or an impulse noise region (block 5), the filter selecting unit selects a filter corresponding to the determined region, and the image data is processed by the selected filter.

Growing community of inventors

Ibaraki, Japan

Yuji Itoh