Abstract: Techniques and tools for coding/decoding of video images, and in particular, B-frames, are described. In one aspect, a video encoder/decoder determines a fraction for a current image in a sequence. The fraction represents an estimated temporal distance position for the current image relative to an interval between a reference images for the current image. The video encoder/decoder processes the fraction along with a motion vector for a first reference image, resulting in a representation of motion (e.g., constant or variable velocity motion) in the current image. Other aspects are also described, including intra B-frames, forward and backward buffers for motion vector prediction, bitplane encoding of direct mode prediction information, multiple motion vector resolutions/interpolation filters for B-frames, proactive dropping of B-frames, and signaling of dropped predicted frames.

Abstract: With intelligent differential quantization, a video codec intelligently quantizes video at differing strength levels within a frame, such as on a macroblock (MB) or a group of MB basis. This allows the codec to control bit usage on a finer granularity than a frame to meet hardware constraints, as well as providing perceptual optimization by coarsely quantizing unimportant regions, while finely quantizing important regions within a frame. The intelligent differential quantization uses motion information gathered from encoding and analysis of the video to classify the importance of different regions of the image, and quantizes the regions accordingly. In addition, the intelligent differential quantization include efficient signaling of information as to the differential quantization strengths in the compressed bit stream.

Abstract: A video encoder performs multi-resolution video coding. For example, the encoder adaptively changes frame sizes to reduce blocking artifacts at low bitrates. A video decoder performs corresponding multi-resolution decoding.

Abstract: Disclosed are reactant feed apparatus for liquid-fueled direct feed fuel cells, including miniaturized versions thereof. More specifically, disclosed is a fuel flow device for delivering liquid fuel to such direct feed fuel cells. The fuel flow device comprises a fuel flow-routing device and an enclosure/partition assembly containing the liquid fuel to be delivered. When the fuel flow device is fluidly connected to the fuel cell, it operates to deliver the liquid fuel to an anode flow field thereof by using the pressurized anodic exhaust gases exiting therefrom as the source of power for pumping the liquid fuel from the enclosure into the fuel cell.

Abstract: The coded block parameters used to code blocks of image samples into structures called macroblocks are compressed more efficiently by exploiting the correlation between chrominance and luminance blocks in each macroblock. In particular, the coded block pattern for chrominance and luminance are combined into a single parameter for the macroblock and jointly coded with a single variable length code. To further enhance coding efficiency, the spatial coherence of coded block patterns can be exploited by using spatial prediction to compute predicted values for coded block pattern parameters.

Abstract: The coded block parameters used to code blocks of image samples into structures called macroblocks are compressed more efficiently by exploiting the correlation between chrominance and luminance blocks in each macroblock. In particular, the coded block pattern for chrominance and luminance are combined into a single parameter for the macroblock and jointly coded with a single variable length code. To further enhance coding efficiency, the spatial coherence of coded block patterns can be exploited by using spatial prediction to compute predicted values for coded block pattern parameters.

Abstract: In an operating liquid feed fuel cell system, fuel concentration in the fuel stream can be calculated as a function of the observed current, the temperature of the fuel stream entering the fuel cell stack, and the temperature of the fuel cell stack itself, thereby eliminating the need for a separate sensor. Typically, methanol will be used as the fuel and the liquid feed fuel cell system will thus be a direct methanol fuel cell system.

Abstract: The coded block parameters used to code blocks of image samples into structures called macroblocks are compressed more efficiently by exploiting the correlation between chrominance and luminance blocks in each macroblock. In particular, the coded block pattern for chrominance and luminance are combined into a single parameter for the macroblock and jointly coded with a single variable length code. To further enhance coding efficiency, the spatial coherence of coded block patterns can be exploited by using spatial prediction to compute predicted values for coded block pattern parameters.

Abstract: In an operating liquid feed fuel cell system, fuel concentration in the fuel stream can be calculated as a function of the observed current, the temperature of the fuel stream entering the fuel cell stack and the temperature of the fuel cell stack itself, thereby eliminating the need for a separate sensor. Typically, methanol will be used as the fuel and the liquid feed fuel cell system will thus be a direct methanol fuel cell system.

Abstract: An encoder dynamically filters information for lossy compression so as to control bitrate or quality with few sudden, dramatic changes to perceptual quality of the compressed information. For example, a video encoder regulates the level of a buffer (e.g., how full or empty the buffer is) by adjusting median filtering of video information (e.g., pixel data and/or prediction residuals). The buffer stores compressed video information for the video encoder. Based upon the buffer level, the video encoder changes the median filter kernel applied to video information. If the buffer starts to get too full, the video encoder increases the size of the kernel, which tends to smooth the video information, introduce slight blurriness, and deplete the buffer. If the buffer starts to get too empty, the video encoder decreases the size of the kernel or stops filtering, which tends to preserve the video information and fill the buffer.

Abstract: The coded block parameters used to code blocks of image samples into structures called macroblocks are compressed more efficiently by exploiting the correlation between chrominance and luminance blocks in each macroblock. In particular, the coded block pattern for chrominance and luminance are combined into a single parameter for the macroblock and jointly coded with a single variable length code. To further enhance coding efficiency, the spatial coherence of coded block patterns can be exploited by using spatial prediction to compute predicted values for coded block pattern parameters.

Abstract: A motion estimation process improves coding efficiency by using a modified search criteria. The modified search criteria takes into account the error signal needed to encode a block of pixels as well as the the motion data when selecting a matching block in a target frame. This approach reduces the combined overhead of both the motion and error signal data for the encoded block of pixels. When used in conjunction with a spiral search path in the target frame, the modified search criteria improves the speed of the search because it eliminates the need for an exhaustive search. A predicted motion vector is used to optimize the search location. Preferably the search order is selected so that target pixels closer to predicted point are searched before pixels farther away in the target frame.

Abstract: To represent the motion of a pixel between successive image frames, this method utilizes the motion information for the transformation block containing the pixel, as well as the motion information for adjacent transformation blocks whenever the adjacent transformation blocks are associated with the same object with which the selected pixel is associated. As a result, transformation errors arising from transformation block discontinuities are decreased and the accuracy and integrity of motion representations of pixels within an object are maintained. Preferably, the motion representations for the transformation blocks are multi-dimensional or affine transformations that are capable of representing complex motions that include any or all of translation, rotation, magnification, and shear.