Here we shortly describe a basis set of terms used in the course. You certainly know them all :-), but human memory is sometimes volatile - it's just for those moments.
Algorithm - a rule (often mathematical) governing computer processes.
Artifacts - Visual effects introduced into a digital image in the course of scanning or compression that do not correspond to the image scanned.
Bandwidth - The transmission capacity of a communications channel, usually expressed in bits or bytes per second (the former is also called baud rate).
Basis - a set of functions that cover a space: i.e., two perpendicular to each other vectors X and Y of lenght one are the Cartesian basis, because we can construct every vector as a linear combination of these basis vectors.
Client-Server Architecture - A systems architecture design that divides functions (which might be a part of a single application) between two or more computers. The client is the machine that requests information; The serner is the machine taht supplies it. A typical client/server architecture for imaging might allow a server to store and transmit a compressed file, and the client to decompress, process, and display the image; see plan "Progressive transmission" in article "Multiresolution analysis".
Compression/Decompression - compression is the process of squeezing more data into a smaller storage space. Decompression is the retrieval of compressed data and its reassembly to resemble its original form (before compression); see also lossless compression, lossy compression.
Decompression - see compression
Derived Image - An image that is created from another image, usually by eliminating part of it. Common techniques used to created a derived image include taking a detail, subsampling to a lower resolution, using lossy compression, or using image-processing techniques to alter an image; also called derivative image.
Detail filter - see high-pass filter.
Fourier transform - transformation that changes a representation of a signal; as a basis we use sine and cosine functions.
Haar wavelet - the simplest wavelet; its scaling function is constant and its mother wavelet is similar to a stair.
High-pass filter - together with low-pass filter must satisfy some special conditions because of exact reconstruction and orthonormality (explanation of this term in article "Mathematical basics"); the wavelet filter is given by the elements of the scaling filter but in backwards order and with every other element negated (see the plan Wavelet transform in "Mathematical basics" article).
Image processing - The alteration or manipulation of images that have been scanned or captured by a digital recording device. Can be used to modify or improve the image by changing its size, color, contrast and brightness, or to compare and analyze images for characteristics that the human eye could not oerceive unaided. This ability to perceive minute variations in color, shape, and relationship has opened up many application for image processing.
Inverse wavelet transform - transformation that changes a representation of a signal; in other words reconstruction; after that we get from wavelet coefficients the original signal, see wavelet transform.
Lossless compression - process that reduces the storage space needed for an image file without loss of data. If a digital image that has undergone lossless compression is decompressed, it will be identical to the digital image before it was compressed. Document images (i.e., in black and white, with a great deal of white space) undergoing lossless compression can often be reduced to one-tenth their original size; continuous-tone images under lossless compression can seldom be reduced to one-half or one-third their original size.
Lossy compression - process that reduces the storage space needed for an image file. If a digital image that has undergone lossy compression is decompressed, it will differ from the image before it was compressed (though this difference may be difficult for the human eye to detect). The most effective lossy-compression algorithms work by discarding information that is not easily perceptable to the human eye.
Low-pass filter - together with high-pass filter must satisfy some special conditions because of exact reconstruction and orthonormality (explanation of this term in article "Mathematical basics"), see high-pass filter.
LZW - Lempel-Ziv-Welch. A proprietary lossless data-compression algorithm.
Mirror filters - two filters that have special coefficients. First filter coefficients are given by the elements of the second filter but in backwards order and with every other element negated; an example of mirror filters can be high-pass and low-pass filter used in decomposition.
Mother wavelet - together with scaling function creates a basis for wavelet transform; it has the same coefficients as a highpass filter used in decomposition.
Noise - data or undentifiable marks picked up in the course of scanning or data transfer that do not correspond to the original.
Pattern recognition - computer-based recognition of forms or shapes within an image.
Pixel - the picture elements that make up an image, similar to grains in a photograph or dots in a half-tone. Each pixel can represent a number of different shades or colors, depending upon how much storage space is allocated for it.
Quality control - Techniques ensuring that high quality is maintained through varios stages of a process. For example, quality comtrol during image capture might include comparing the scanned image to the original and then adjusting colors.
Resolution, image - Number of pixels (in both height and width) making up an image . The higher the resolution of an image, the greater its clarity and definition.
Resolution, output - The number of dots per inch, dpi, used to display an image on a display device (monitor) or in print.
Scaling filter - see low-pass filter
Scaling function - together with mother wavelet creates a basis for wavelet transform;it has the same coefficients as a lowpass filter used in decomposition.
Smoothing filter - see low-pass filter
Subsampling - Using an algorithm to derive a lower-resolution digital image from a higher-resolution image (i.e., elimination every other pixel in each direction); see derived image.
Vector graphic - a digital image encoded as formulas that represent lines and curves.
Wavelet - a function that "waves"; a set of wavelets creates a basis of a wavelet transform; we can distinguish a lot of kind of wavelets: Haar, Mortlet, Daubeshies, etc. wavelets; they look different and they have different properties: can be orthogonal, bi-orthogonal, normalized etc.
Wavelet filter - see high-pass filter
Wavelet transform - transformation that changes a representation of asignal; in other words decomposition; after that we get some coefficients that completely and uniquely represent the signal; a basis for this transformation is set of wavelets; see also inverse wavelet transform.
Quadrature filters - see mirror filters.