Last time blog post (based on HTML5 's PACS--HTML5 image processing (7) to realize the Client JS window width window), we introduce how to use the dicom pixel data parsed in the background to generate the image to be displayed and adjust the window width in the JS end. Several problems are encountered:
1. for (2050,0020) Presentation LUT shape equals inverse or (0028,0004) photometric interpretation equals MONOCHROME1 grayscale image, Using the lookup table algorithm in the text, the image that comes out is actually the inverse color image. To make changes to the lookup table, increase whether the color is reversed, and when the pixel is generated, it is modified according to this tag.
2. for (0028,1040) Pixel Intensity relationship equals Lin and (0028,1041) Pixel Intensity relationship sign equals-1 grayscale image, to preprocess the pixel data , otherwise out of the image may be completely wrong. Refer to the DICOM standard c.8.11.3.1.2 instructions.
3. The 2nd problem came out of the rough image, once as I mistakenly think that the large image of DX to use a non-linear window width adjustment algorithm, halfway to a friend to ask and to a few open source dicom look at the tool research, found that the use of the window width of the linear algorithm.
4. js in the large array of address lookup speed is a bottleneck, you can work to improve the speed of image generation.
For the record, first ...
Use Radiand DICOM view to see the above DX image, and adjust to the same window width (image effect has no).