September 02,2010
CEDIA 2010: Booth# 5015 Read More »
March 22,2010
NAB 2010: Booth# SL10723 Read More »
January 27,2010
Dual De-interlacer/Scaler with Analog Front-End Read More »
TECHNOLOGY
We welcome your suggestions and comments about this section at : info@i-chipstech.com
Glossary of Terms
A/D Converter (ADC)
Device located on the input side of a circuit for converting the video signal from analog to digital. An ADC for digitizing video must have a very high sampling rate.
Active Video
The part of a video signal that is visible on the screen, i.e. outside of the blanking intervals.
AOI (Area Of Interest)
Portion of an image used for processing. On the i-Chips devices, the AOI can be defined in 2 ways: 1) the raster scan coordinates of the upper-left and bottom-right corners, 2) the raster scan coordinates of the upper-left corner and the horizontal and vertical widths of the region
Aspect Ratio
The measurement of a viewing area in terms of relative width and height. For most current displays, this ratio is 4:3. For HDTV, the ratio is 16:9.
Binarization
Process of converting a gray-scale image to a binary image. Refer to the definition of Thresholding.
Blanking
In a raster scan image, the time during which the video signal is suppressed to allow for the retrace. There are blanking periods both in the horizontal and vertical directions, respectively called HB and VB. For most video formats, the blanking periods amount to about 25%-30% of the signal bandwidth.
Blob
A single connected region in a binary or gray-scale image.
Blob Analysis
Identification of the segmented objects in an image based on their geometric features such as area, length, etc. The process used to locate blobs is called Labeling. For a complete description of the geometric features computed by the i-Chips devices, please refer to the datasheets of the IP90C18 and IP00C101.
Component Video
The separation of chrominance (color) and luminance parts of the video signal. In component video, these two signals are recorded separately, which helps maintain better picture quality over more generations.
Composite Video
The complete visual wave form of the video signal composed of chrominance and luminance picture information, blanking pedestal, field, line, sync pulses and field equalizing pulses.
Convolution
Also called a spatial filter, the convolution is a sum-of-products calculation in a local area of the image. At each position of the local area, the pixel value is multiplied by an operator (filter coefficient), and the summed result is used to replace the value of the center pixel. A typical filter would make use of a 3x3 or 5x5 local area. This computation is usually performed on all the pixels of the image to produce a “filtered” image. There are a number of proven filters (or sets of coefficients) that can be used for example to detect edges, enhance contrast or to remove noise.
EDO DRAM
Abbreviation of Extended Data Output Dynamic Random Access Memory, a type of DRAM that is faster than conventional DRAM. Unlike conventional DRAM, which can only access one block of data at a time, EDO DRAM can start fetching the next block of memory at the same time that it sends the previous block to the CPU.
DRAM
Acronym for Dynamic Random Access Memory, a type of memory used in most personal computers. The term dynamic indicates that the memory must be constantly refreshed or else it will lose its contents. See also SDRAM, EDO DRAM and SGRAM.
DVI
Acronym for Digital Visual Interface, a specification developed by the Digital Display Working Group to provide a high-speed digital connection for visual data types that is display technology independent.
The interface is primarily focused at providing a connection between a computer and its display, and its specification aims at uniting the different industry segments around one monitor interface standard.
De-interlacing
Process by which an interlaced scanned image is converted to a progressive scanned image. De-interlacing can be achieved by a number of different methods, and with various levels of quality of the output image. The most simple technique, called “field merging”, recombines the odd and even fields together to create the output image. This technique will work well for static images, but produces image artifacts when working on video images where there is motion in the image. This is caused by the fact that there is a difference of 1/60th of a second (for an NTSC image) between the odd and even fields, and an object in motion will not be at the same exact same position in the two fields. De-interlacing based on “field interpolation” produces a better quality image, because each field is individually scaled up to fit the output resolution of the progressive scanned image.
Dithering
Process used when the color palette available is smaller than the color content of the image. For example, when converting from 24-bit color to 8-bit color (an 8-bit palette has only 256 colors compared to the 16.7 million colors of the 24-bit palette), dithering uses a combination of the available colors to simulate the original color of the pixel in the image. Dithering is also known as "error diffusion."
The i-Chips devices support 2 modes of error diffusion when converting from 10-bit color to 8-bit color, or from 8-bit to 6-bit. In the intra-frame (spatial) mode, the fractional part of the pixel value is carried over the neighboring pixel in the image. In the inter-frame (temporal) mode, the fractional part is carried over the same pixel in the next frame.
Feature Extraction
Process of determining the features of an image. Typical features include histogram, X and Y projections, and region characteristics such as area, perimeter, center of gravity, etc.
Field
The set of alternating lines in an interlaced video frame. An interlaced video frame consists of two fields – odd and even. Each field is one-half of a complete scanning cycle (1/60 of a second in NTSC; 1/50 of a second in PAL/SECAM), and is made of either the odd lines or even lines. When interlaced, two fields combine to make one video frame.
Field/Frame Synchronization
Process that uses the synchronization signals to eliminate the problem of “frame tearing” when performing scan rate conversion. On the i-Chips SCREEN devices, this is accomplished by aligning the output field or frame on the input timing with an adjustable delay.
Field Frequency
The rate at which a complete field is scanned or displayed, normally 59.94 times per second in NTSC.
Frame
A single, complete picture in a video or film recording. A video frame consists of two interlaced fields of either 525 lines (NTSC) or 625 lines (PAL/SECAM), running at 30 frames per second (NTSC) or 25 fps (PAL/SECAM). Film runs at 24 fps.
Frame Buffer
Name of the memory device used to store, pixel by pixel, the contents of an image.
Frame Rate Conversion (FRC)
Process of changing the frame rate of a video signal. Because the output frame rate is different from the input, a memory is required for the temporary storage of the video data during the conversion process. The i-Chips devices use the technique of double buffering and read/write pointer management to ensure consistent output image quality.
Gamma
Characteristic of CRT displays and cameras that quantifies the non-linear relationship between the output brightness and the input voltage. Gamma compresses small signals so that the contrast in the low luminance will be lost. The gamma numerical value can range from 1 to 5. For design purposes, NTSC standardized gamma at 2.2. PAL and SECAM standardized gamma at 2.8.
Gamma Correction
Process used to compensate for the gamma of the display or camera, or, in other words, to restore the linear relationship between brightness and voltage. Gamma correction is numerically equal to the inverse of the display gamma, and because it is non-linear, it can be best achieved with a look-up table for each color.
Gray-scale Image
Image that is digitized with a fixed number of gray levels. For 8-bit data, the image contains 256 (28=256) possible shades of gray, which is actually more shades of gray than the human eye can see.
HDTV
High-Definition Television. The proposed standards will approximately double the horizontal and vertical resolutions, and increase the screen aspect ratio from 4:3 to 16:9. The most common formats are:
1080i 1080 lines, 1920 pixels, 30 frames per second, interlaced scanned.
720p 720 lines, 1280 pixels, 24, 30 or 60 frames per second, progressively scanned.
Histogram
Graphical representation of the frequency of occurrence of each intensity (gray level) of the pixels in an image. The vertical axis represents the number of times each intensity is recorded. A histogram skewed heavily to the left indicates a dark image, while a histogram skewed to the right indicates a bright image. The histogram is often used to determine an appropriate value for thresholding an image.
JBIG
Compression standard that uses arithmetic encoding to compress bitonal images very efficiently. In 1993, the International Organization for Standardization (ISO) adopted JBIG as an international standard, ISO/IEC Standard 11544. For bitonal images, the JBIG method provides higher compression ratios than the conventional MH, MR and MMR techniques. As a lossless method, it allows to reconstruct the original image perfectly.
Keystone Correction
Process of compensating for the image distortion caused by the vertical projection angle. Mainly used in projectors, the keystone correction converts the image to a trapezoidal shape before being output to the optical lens system. This ensures that the projected image will retain its original rectangular shape. Keystone correction on the i-Chips devices is fully programmable.
Lossless Compression
A compression technique that preserves all the original information in an image or other data structures.
LSI
Abbreviation for “Large Scale Integrated” circuit, used to designate a chip or device that generally has more than 1,000 and less than 10,000 transistors.
LUT
Acronym for Look-Up Table, which is a high-speed memory used to “map” each pixel of the input image to a new value. For example, the LUT is used to generate inverse video or pseudo-color images.
Machine Vision
The use of devices for optical non-contact sensing to automatically receive and interpret an image of a real scene, in order to obtain information and/or control machines or processes.
Median Filter
Method of image filtering which replaces each pixel value by the median gray-scale value of its immediate neighbors. When implemented in software, the median filter is computationally expensive, because it requires sorting all the pixel values in the region. The median filter is a powerful tool to help eliminate high-frequency noise in the image, while preserving the quality of the edges.
NTSC
Acronym for the National Television Systems Committee of the Electronics Industries Association (EAI).
The NTSC prepared the specifications for U.S. commercial color broadcasting approved by the Federal Communications Commission in December 1953.
NTSC Format
A color television format having 525 scan lines, a field frequency of 60 Hz, a broadcast bandwidth of 4 MHz, line frequency of 15.75 kHz, frame frequency of 1/30 of a second, and a color subcarrier frequency of 3.58 MHz. NTSC uses YIQ.
OSD (On-Screen Display)
Process of superimposing text on an image. For example, OSD is used on a TV to display a menu for system adjustments or to display text for hearing-impaired persons.
Panorama Conversion
Function of enlarging an image horizontally in a non-linear fashion by applying a ratio that increases as the ends of the screen are approached. This is necessary when converting images with 4:3 aspect ratio to a wide-screen 16:9 output. The i-Chips SCREEN devices support non-linear zooming for panorama conversion.
PiP (Picture in Picture)
Term used to identify a small image window that is displayed within a larger picture. This is a feature on some display devices that allows the viewer to watch video signals from two different sources simultaneously.
Projection (X- or Y-)
The X-Projection is the sum of all the gray-scale values in each column of the image. The Y-Projection is the sum of all the gray-scale values in each row of the image. Projections are useful to detect the presence or absence of an object in a scene, or to locate a blob or object.
RGB
Acronym for Red-Green-Blue used to define the three 8-bit color channels of an image. When combined, the red, green and blue channels provide up to 16.7 million possible combinations, hence colors. The 24-bit encoding of the color is also referred to as “True Color”.
Resolution
When referring to an image or a display device, this term defines the total number of pixels, or dots, that can be addressed. This number is expressed in terms of a column (pixels) by row (lines) figure. The higher the resolution the more pixels that can be displayed on a screen. The following resolution standards have been defined:
Standard Resolution Description
VGA 640 x 480 Video Graphics Array
SVGA 800 x 600 Super VGA = 1.25 x VGA
XGA 1024 x 768 Extended Graphics Array = 1.6 x VGA
SXGA 1280 x 1024 Super XGA
UXGA 1600 x 1200 Ultra XGA = 1.5625 x XGA
QXGA 2048 x 1536 Quad XGA = 4 x XGA
QSXGA 2560 x 2048 Quad SXGA = 4 x SXGA
Run Length Encoding
Data reduction method to code a binary image. For each line in an image, data is stored denoting only the starting location of a blob (region) and the length of the run of that line over the region.
SDRAM
Abbreviation of Synchronous DRAM, a type of DRAM that can run at much higher clock speeds than conventional DRAM. Synchronous control means that the DRAM latches the read and write requests from the CPU, allowing the latter to perform other tasks.
SGRAM
Abbreviation of Synchronous Graphic Random Access Memory, a type of DRAM used on video adapters and graphics accelerators. Like SDRAM, SGRAM can synchronize itself with the CPU bus clock up to speeds of 100 MHz. In addition, SGRAM uses several other techniques, such as masked writes and block writes, to increase bandwidth for graphics-intensive functions.
Signal-to-noise (S/N) Ratio
The strength of a video and/or audio signal in relation to noise. The higher the S/N ratio, the better the quality of the signal.
Smart Camera
New term for a complete vision system contained in the camera body itself, including the imaging device, image processing and decision making functions. Mainly used in industrial inspection, smart cameras do not need a PC for processing the images: they output a signal to either accept or reject the part being inspected.
S-video
Type of video signal that transmit the luminance and color portions separately, using multiple wires, thus avoiding the NTSC encoding process and its inevitable loss of picture quality. Also known as Y/C video.
Template
Artificial model of an object (or feature of an object) represented by a matrix of pixel values. The Template Matching Processor IP90C08 supports configurable template sizes (112x1, 32x3, 16x7, etc.).
Template Matching
Process by which the template is compared to the actual image to determine how well they match. Template matching on the IP90C08 is done by computing the sum-of-differences in gray-scale values between the template image and the pixels in the input image.
Thresholding
Process of converting a gray-scale image to a binary image. If the pixel value (intensity) is above the threshold, it is converted to white, and if below the threshold, it is converted to black. The IP00C101 device supports two modes of thresholding. In the normal mode, the user defines a high and low threshold to allow clipping of the center portion of the histogram. In the hysteresis mode, the output pixel value is computed from the intensity of the pixel both in the current image and in the previous image.
VESA
Acronym for the Video Electronics Standards Association, a consortium of video adapter and monitor manufacturers whose goal is to standardize video protocols.
Y/C Separator
Circuitry used in a video decoder to separate the luma and chroma in a composite video signal.
YCbCr
Color space defined by Recommendation ITU-R BT.601. Y is the luma component (luminance) and the Cb and Cr components are color difference signals (chrominance). Cb and Cr are scaled versions of U and V in the YUV color space.
4:4:4 YCbCr means that for every sample of Y, there is one sample each of Cb and Cr.
4:2:2 YCbCr means that for every two samples of Y, there is one sample each of Cb and Cr.
4:1:1 YCbCr means that for every four samples of Y, there is one sample each of Cb and Cr.
Many people use the YCbCr notation rather than Y'CbCr or Y'Cb'Cr'. The technically correct notation is Y'Cb'Cr' since all three components are derived from R'G'B'.
YPbPr
Color space used when three video components are to be conveyed in three separate channels with identical unity excursions. Y is the luma component and Pb and Pr are the color difference components. YPbPr is employed by analog component video equipment such as M-II and BetaCam. Pb and Pr bandwidth is half that of luma.
YUV
Color space used in composite NTSC, PAL or S-Video systems. Y is the luma component and U and V are the color difference components. Many people use the YUV notation when they actually mean YCbCr data. The U and V components are typically modulated into a chrominance (chroma) component:
C = U*cos(t) + V*sin(t)
Where t represents the ~3.58 MHz NTSC color sub-carrier.
The two signals Y and C can be conveyed separately across an S-Video interface, or Y and C can be combined into composite NTSC or PAL. U and V are only appropriate for composite transmission as 1-wire NTSC or PAL, or 2-wire S-Video. The UV scaling is incorrect when the signal is conveyed as three separate components.
YUV to RGB Conversion
Conversion supported by the i-Chips devices when handling input video signals in YUV format. The conversion is carried out with coefficients that can be programmed in the device. These coefficients are fixed for each type of video signal, i.e. NTSC, PAL, etc.
