DICOM standards and applications-lecture 5 information organization and performance of Medical Images

As we described earlier, DICOM is a standard for medical images. The previous section describes the exchange of images in storage media and communication environments. This section describes the organization and performance of images.
Presentation refers to the process of visualization of image data on display devices or films. It requires consistent display effects on devices of different systems and features, that is, visual equivalence. In this way, clinical application requirements can be ensured.
To this end, the DICOM Standard specifies the corresponding image information organization and processing functions. The following is an introduction.
1. image encoding format
DICOM uses bitmap to describe the image. That is, the color, brightness, and other information at the position are displayed point by point. Only brightness information is available for a monochrome image, which is called grayscale. Different color representation methods exist for color images. Generally, the three primary colors of RGB are used, that is, a vertex is represented by the values of red, green, and blue. DICOM allows three matrices (Bit Planes) to represent three components respectively, and only one matrix to represent the entire image. In this case, each point in the matrix is composed of three values.
For a pixel value, DICOM is called the sample value ). The description method of the sample value is given by three data elements. Bits allocated indicates the binary digits stored in the sample value. Bits stored refers to the number of actually occupied bits. High Bit indicates the position of the highest bit in the allocated storage unit.
For example, if the pixel storage format of an image in a CT is 1, the allocated bits is 16 bits, the number of BITs is 12, and the maximum bits is 11.
It must be noted that, in actual access to image data, the actual storage unit address order of high and low pixels must also be determined by the big-Endian or littleendian attribute in the transmission syntax. Of course, there are many other attributes of the image in the DICOM standard. You can refer to the specific content of the standard.
DICOM supports multi-frame images by encapsulating multi-frame images in a pixel data element. It is specified by the number of frames attribute.
Ii. Brief description of compression methods
Raw medical images occupy a large amount of storage and are less efficient during transmission and storage. Compression must be used to reduce redundant information in images, to reduce the number of bytes required for image storage without loss of image information or loss, it is necessary to shorten the communication transmission time and reduce the storage space.
Compression methods include lossless compression and lossy compression. The lossless compression method can restore the original data intact, while lossy compression is irreversible and cannot recover to the original situation. Lossless compression due to restoration constraints, the compression is relatively small, generally 2 ~ 10: 1. However, lossy compression can achieve a large compression ratio, typically 10 to 10 ~ 200: 1, or even 300: 1. Based on the characteristics of the compressed object, lossless compression is suitable for text compression, because the original meaning cannot be expressed after the text information is lost. However, lossy compression is suitable for multimedia information such as voice and image. This information is felt by human auditory and visual organs and is highly redundant, there is no damage to the understanding of information due to proper loss, and a high compression ratio can be achieved.
Lossless Compression mainly uses run length encode (RLE) and Huffman encoding. Lossy compression is commonly used to convert and compress information into another representation domain. Redundancy is removed by using the distribution characteristics in different domains. For example, if the image is converted to the frequency field through ffu's transformation, the image's low-frequency part is saved, and some details are lost for compression. For example, if the image is transformed to different scales through wavelet transform, and the transformed values in different regions under different scales are saved, a high compression ratio can be achieved.
Due to Medical Liability and legal reasons, the western medical community has taken a very cautious attitude towards lossy compression of medical images. Reflected in DICOM standard, lossless compression is recommended. Specifically, we use a simple RLE and JPEG standard lossless compression algorithm.
There are many forms of RLE for program encoding. The Encoding Algorithm Used in DICOM is as follows:
Replace duplicate bytes with <-number of bytes + 1> <byte value>.
For non-repeated bytes, use <number of bytes-1> <Non-repeated byte sequence> instead.
Image Format 2 encapsulated by RLE compression.
JPEG is currently the most widely used image compression standard. The JPEG standard contains multiple methods of lossy compression and lossless compression. The basic process of JPEG compression is to divide the compressed image into 8x8 squares. First, perform the differential encoding to reduce the code length, and then perform lossless compression using Hoffmann or arithmetic encoding, or use discrete cosine encoding for lossy compression. The JPEG standard has a high compression ratio and a good effect. Due to the complexity of the algorithm, I will not provide a detailed description here. For details, refer to the relevant literature. The unique identifier of DICOM with JPEG lossless compression is "1.2.840.10008.1.2.4.70 ".
Three Gray Display
Digital signals from images can be accurately and purposefully measured, described, transmitted, and reconstructed. However, visual interpretation of signals relies on systems that display different features of an image. Therefore, images generated from the same signal may have different visual representations, information, and features under different display devices.
In medical images, it is important to maintain consistency when a variety of digital images are presented, regardless of whether the image is observed. For example, they should be displayed on the video monitor of a workstation or played as slides. If there is a lack of specifications that cannot be formed in the visual representation of different devices, it is very likely that when an image is observed under a certain device, it has very good diagnostic value, but it is very different from the former when observing on another device, which greatly reduces the diagnostic value. Therefore, DICOM provides a purposeful and high-quality mechanism to describe digital image values to different levels of brightness values, if this method is applied to the association between known digital image values and display brightness, display on display devices with different properties can produce better visual consistency. The relationship between digital image values and display brightness defined by DICOM is a model and measurement standard based on human's understanding of a large range of brightness, it is not based on any image display device or the morphological feature description of any image format, nor does it depend on your personal preferences. Because it is easy to be called correctly by other structures such as DICOM query tables. The DICOM standard uses the BARTEN model of the human visual system to process image data and achieve visual consistency.
DICOM defines the standard gray level display function in the standard image system from the mathematical point of view. These image systems can be hard-copy printers or electronic display systems used to display soft-copy content. Hard copy includes both transparent and reflective printed images. The images in these prints have different optical density in the image due to the influence of scattering reflection during transmission. For an observer, each pixel with a certain brightness in the image it sees is determined by the illumination and light density of the image elements. Soft copies are produced by emission display systems (such as CRT monitors) and electronic optical valve systems (such as light source and liquid crystal display systems.
The purpose of DICOM is to display the image by using digital-driven layers to produce brightness and light density changes. The predictability of image transmission applications, such as morphology, valuable values (VOL), and LUT Lookup tables, all need to know the characteristic curves of the display system. If you standardize the response functions and curves of the display system, it makes it easy to transmit images in different display systems (such as in a network environment.
Figures 3 and 4 illustrate the content of the standard grayscale display function, which is part of the image performance. Before using the grayscale display function, there are many other steps for image modification. The image acquisition device will make appropriate adjustments when the image is generated. Other links will generate some windows or layers to select the dynamic range of each part for the image performance. Even if other links can be used to adjust and display the dynamic regions selected in preparation. Displays the LUT output P value (expressed value) of the query table ). These P values become the digital driver level (DDL) in the standard display system ). The standard grayscale display function maps the P value to the logarithm of the brightness in the standard display system. The ing process generated by the standard display system is independently completed by the system.
When the DICOM model, display chain, and standardized display system acquired by images are represented by P values, the devices tend to be independent and visually linear. In other words, regardless of the performance of the standardized display system, the same P value range will be very similar.
The main objective of DICOM is to define a class of appropriate standard grayscale display functions for all image presentation systems from a mathematical perspective. The purpose of defining this type of standard grayscale display function is to allow you to explore how previous P values are converted to the observed brightness values in a standardized display system. In essence, the standard display function defines the units of P value output from LUT display, and can also be used as a digital driving level for the standardized display system.
The second goal of DICOM is to select a display function, this function provides a gray-level perception or basic representation of the digital driving levels of a given image at different brightness levels and an easily usable and measurable display system. several similar layers. While many other functions can serve the first goal, this grayscale display function can be selected to achieve the second goal. For this type of function, the P value is almost linearly related to the human perception response. Similar but not guaranteed to have the same information. A display system with a greater brightness area range and a higher brightness will be able to show more nuances.
Because the perception is determined by two factors: the image content and the observer's subjective thoughts, the Brightness Difference is for an observer, similarly, we cannot strictly follow the absolute linearity of perception. To achieve absolute linearity awareness, the image performance must be adjusted by certain methods defined in the DICOM standard (for example, vol and LUT display) to meet user expectations. Without this defined display function, it will become very complicated when you encounter adjustment problems for such display systems with a wider range and more types on the network, processing is also relatively difficult.
The characteristic curves of the display system are generally different from those of the standard grayscale display function. These devices may include the merging method, especially those methods defined in the transformation that can make the device consistent with the grayscale standard display function. DICOM provides instances for testing these display systems. These display systems generally have such a feature that their performance work can be scalar, and the process is similar to the evaluation of the standard grayscale display function.
DICOM does not specify specific display functions for color images.
Conclusion
The image is the core of DICOM standards. In addition to the previously introduced image storage and image transmission, the image performance is also a very important part. It involves the user's final feelings about the image, this affects Image Understanding and disease diagnosis.