KR100629808B1 - Scaleable resolution motion image recording and storage system - Google Patents

Abstract

A system for scaling the resolution of an image input stream to a desired resolution is provided. The system includes a processor for preprocessing an image input stream; An IO system for receiving an image input stream; An input analog converter for converting an image input stream, a frame buffer for storing image frames, a router for decorrelating the image input stream among one or more processors, a time processor for temporal conversion of the image input stream, an encoder A DMA channel for delivering the acquired data to the storage device; And a local storage device for storing the processed image input stream. This method uses the described system components to generate a scalable resolution motion image record that can be stored and reproduced on any number of other playback devices.

Decorrelation, DMA Channels, Router, Local Storage, Input Analog Converter, Stream

Description

SCALEABLE RESOLUTION MOTION IMAGE RECORDING AND STORAGE SYSTEM}

The present invention relates to the field of digital signal compression and quantification. More specifically, the present invention relates to systems and methods for providing recording and storage of scalable resolution images.

Electronic motion image recorders have traditionally been designed in one or at most only a few motion image formats. The emergence of various high definition video, medical, scientific, and industrial formats has created a need for single systems that can support a wide range of motion image formats. As motion image frame rate, image size, and pixel resolutions have increased, the amount of data that must be maintained to represent a motion image stream in the sample domain has increased tremendously. This puts a heavy burden on processing, storage, and communication costs to support these data sets with the desired resolution.

The present invention provides an efficient and cost effective system configured to effectively support all motion image formats. The system can be modularly extended to increase throughput and trade off system throughput between the number of image streams, their frame rates, frame resolution, and pixel resolution. The system uses subband methods to fully support variable image size and frame rate recording. This allows dividing a single image stream into multiple lower rate streams to allow more reasonable processing speeds. The present invention also enables applications for variable sized arrays of standardized image processing components.

More specifically, optimized storage and efficient communication are achieved by storing image streams in the information domain rather than the sample domain. This generally provides a significant reduction in storage and communication conditions, while still providing guaranteed recording quality. Unlike conventional video tape recorders, this system can write to the same removable storage medium with any desired image resolution, frame rate, and quality.

The following description illustrates the method of the present invention carried out by the system of the present invention. However, this method can also be implemented by other image processing apparatus.

The following provides a list of components that may be included in the system of the present invention. The functions accomplished by each component are also provided. As will be appreciated by those skilled in the art, this list is not exhaustive and other components that provide similar functionality, such as other band limiters, band splitters or color space conversion modules, may also be added.

Ingredient function

-------------------------------------------------- ------------

Input Bandsplit Input Converters and Optional Analog Bandsplit

Deframer Digital Data Stream In

Color Space Optional RGB → CbYCr

Band limit 1D or 2D band limit and any genera of components

                                 Degree conversion

Band Split Full Band 2 pixel color 1/2 band saturation dual

                                 To the pixel "quad"

CbYCr → CbLCrH

                                 Full Band 4 Pixel Colors into Visual Space "Quad"

CbYCr → CbLCrD

             Solid color to band split "quad"

MM → --L--H

             Solid to 2D Band Split "Quad"

MMMM → LHVD

Sample only part of an inverse pulldown input field or frame

Framebuffer Buffer and Optional Time Multiplex Frames

Interface Processing Interlaced Image Processing, Interlaced Teeth                 

                                 Allowing unknowns to be treated as progressive

                                 Will

Route router N streams to N IP channels

                                 Use time multiplexed frames

                                 Simple temporary band split of frames

                                 Used

Temporary IP Temporary Conversion and Time Multiplication of Optional Frames

                                 Flex

Interlaced Processing Interlaced Image Processing, Interlaced Teeth

                                 Allow image to be processed as progressive

                                 To do

Spatial IP Spatial Transformation

Quantification Required Ingredients and Temporal Quantification

Entropy coding entropy coding, potential respective temporal and spatial

                                About ingredient

DMA channels provide a channel for data movement

RAM Queues Peak Buffering

Disk Queues Peak Buffering

Disk, Tape, or Channel Final Output

IO subsystems accept motion image input streams in one of various pixel formats.

RGBA red, green, blue and alpha

YCbCrA roughness, blue color difference, red color difference, and alpha

YC roughness and alternate color difference

M solid color

MM two consecutive monochrome samples

MMMM Four consecutive monochrome samples

The input analog converter can operate at a fixed rate, converting up to N times per second, where N is determined by the particular implementation. If more than N conversion rates are required, the input system can use an analog band split into 2 or 4 streams at 1/2 or 1/4 speeds of the input signal to allow more than N multichannel conversions. .

In one embodiment, the digital IO system converts these input pixel streams into quad component representation streams. In each case, it is assumed that the components are decorrelational and can be processed separately. However, if the components are not decorrelated, a color space transform can be used to decorate the color components. This also includes subbanding certain color components to further correlate the components to spread the bandwidth evenly among the available channels. Although splitting the pixel stream into quad component representations is described as one embodiment, image processing resources can also process the pixel stream by splitting the pixel stream into a number of different component representations. For example, the system can be used to process dual channel image streams. In these cases, the first and third components are assigned to the first channel and the second and fourth components are assigned to the second channel.

CbYCrA Single Color Pixel, 4 Component Full Band

CbLCrH dual color pixels, 2 subband luma components (L and H), 1/2 band monochrome lu

         Harr transform odd near orthagonal 7

         Band-split using a tap filter or an odd near quadrature 9 tap filter.

CbLCrD 4 color pixels, 2 1/2 band luma components (L and D), 1/4 band solid color.

        Luma is half the original bandwidth, but is limited to a diagonal rectangle that provides full horizontal and vertical response. It is band-split into half lower and half half using a 2D filter of 3x3 or 9x9 taps. Color is band limited to half band in two dimensions.

LMMH Quad Monochrome Pixel, 1D 4 Subband Components

LHVD Quad Monochrome Pixel, 2D 4 Subband Components

Frame buffers store 1 to N frames of the input stream and distribute 1 to N frames to 1 to N signal processors. This enables support of temporal processing in time multiplexing image processing resources, or frame buffers or routers. N is determined by dividing the required quad throughput by the quad throughput of the respective signal processors. This also allows the support of temporal processing in the frame buffer or router.                 

The router may achieve simple spatial or temporal band splits. This can lead to additional decorrelation and allows subdivision of the signal stream among computing resources. It can also be used to support image resolution or throughput over physical buffers or individual throughput supported by image processing modules.

This system can also perform temporal conversion processing. In the preferred embodiment, a temporal processor is needed for each quad stream that needs processing. The input to the temporal processor is a stream of image frames. The output is a stream of shuffled time transform frames. For example, if one temporal transform is selected, an alternate stream of low and high temporal component frames is created. If two temporal transforms are selected, a stream of low, midlow, midhigh, and high temporal component frames is generated. Temporal transforms can be inserted between the router and the spatial transform processor as desired.

Spatial signal processors from 1 to N accept quad streams of frames. These processors perform spatial multiband conversion, quantify data for a particular recorded signal quality and encode entropy. DMA channels 4 through 4N deliver the encoded data to a RAM buffer that coordinates peak transfer loads.

This data can then be transferred to standby or output files on disk, tape or other peripherals capable of storing digital information. If the data rate is higher than what a particular peripheral can support, this information will wait in faster storage until the desired peripheral can accept this data. In the case of a fixed speed peak tolerant recording, a phase delay is specified, and the hierarchy of processes related to the hierarchy of buffer devices monitors the speed and potentially controls the quality level. This makes the process as acceptable for the peak as possible. Each process responsible for buffering determines whether data bundles are exceeded for the time scale for this processor . If the local average speed exceeds the set point, the quality level of the images being received by the output or storage device is reduced until the local average speed no longer exceeds the set point. If the average is lower than the set point and the quality level is lower than the target, the quality level of the images increases.

<Example>

This is an example of a unique storage process. The target data rate is 12MB / S and the quality level is 66dB. This causes a high peak rate of 80 megabytes per second to occur for one frame time, a high peak rate of 40 megabytes for about one second, and a peak rate of 20 megabytes per second for a few seconds. Stepped up or down is added together to the goal to get the next level of quality. The quality level is recalculated for every frame.

Process Time Scale Set Point Peak Rate dB Steps

----------------------------------------------

4 frames 40 MB / S 80 MB / S 3 dB for RAM

4 seconds to disk 20 MB / S 40 MB / S 1 dB

30 seconds to tape / channel 10 MB / S 12 MB / S 1/2 dB                 

The playback process is the reverse of the recording process, except that playback from a peripheral that is slower than required will result in preroll time and a faster peripheral is used to wait for the required amount of data. Preroll time is the time required to transfer data from a slower peripheral to faster storage. For example, if data is stored on a compact disc, this data can be prerolled in RAM to provide playback at the appropriate frame rate and quality level.

Although the foregoing description has many details, these descriptions should not be construed as limiting the scope of the invention, but merely as providing some examples of presently preferred embodiments of the invention. For example, although this method has been described with reference to standard motion and still images, this method is used to optimize the quantification of any signal stream. The scope of the invention is also to be determined by the appended claims and their legal equivalents, rather than by the examples given.

Claims (14)

A system for scaling resolution of an image input stream, A processor for preprocessing the image input stream; An IO system coupled with the processor for receiving the image input stream; An input analog converter coupled with the IO system; A frame buffer, coupled with the input analog converter, for storing image frames; A router coupled to the input analog converter, for correlating and dividing the image input stream among one or more processors; A temporal processor, coupled with the router, for temporal conversion of the image input stream; A DMA channel, coupled with the temporal processor, for delivering encoded data to a storage device; Local storage device for storing the processed image input streams associated with the DMA channel System comprising a. The method of claim 1, The input analog converter further comprises an analog band splitter for dividing the image input stream into multiple streams. The method of claim 1, The IO system further comprises a color space converter for decorrelating color components. The method of claim 1, The system includes four DMA channels. The method of claim 1, The system includes one or more time processors. The method of claim 1, The system further comprises a time conversion module coupled to the router and the space conversion processor. A method for processing an image input stream, Receiving a desired image input stream conversion rate; Converting the image input stream into one or more presentation streams; Calculating a frame conversion rate; In response to the frame conversion rate being lower than the desired conversion rate, and in response to the input stream components being correlated, converting the color space of the image input stream and dividing the image input stream into one or more representation streams; Routing the image streams to one or more signal processors; Performing a time conversion process on the representation streams; Storing post time processing of the presentation streams How to include. The method of claim 7, wherein Splitting the image input stream into two representation streams comprises splitting the image stream into four quad component representations. The method of claim 7, wherein Splitting the image input stream into two representation streams comprises performing an analog bandsplit on the stream. The method of claim 7, wherein And performing time conversion processing on the representation streams comprises processing the primary representation stream using a separate time processor. The method of claim 7, wherein Performing temporal transform processing on the representation streams comprises performing spatial multiband transformation of the stream. The method of claim 11, And performing a time conversion process on the representation streams comprises quantifying the data in the stream. The method of claim 12, And performing time transform processing on the representation streams includes entropy encoding the stream. The method of claim 7, wherein Storing the streams, Calculating a data bundle of a first storage device used to store the data; In response to the scrim requesting a higher data bundle than is supported by the first storage device, waiting the data on a second storage device until the first storage device can accept the stream. How to include.