Fortunately, the (uncompressed) MetaImage disk format was so straightforward even I could understand it, and it was even suprisingly performant. I wanted to have a thin and pure Python wrapper around numpy that would allows me to read in and write out image data. ITK's Python bindings (SimpleITK) was not pippable or easily usable yet, and I found working with image data as numpy arrays far preferable and faster than using ITK as a library in custom C++ programs which I'd need to compile and recompile as an analysis developed. This project started out at a time when I was analyzing lots of Gate image outputs. SimpleITK write also only seems to produce usable dicoms files when updating an existing image, not when creating a new one from scratch. If it would, it would require SimpleITK, primarily because pydicom does not support dicom image write. This component is governed by its own license.ĭicom write is not supported right now. For NKI decompression I supply a 64bit Linux and Windows lib, if you need support for other platforms you can compile the function in medimage/nki_decomp yourself. Of particular interest perhaps are the DVH analysis function, and the distance to agreement calculation. Included are some basic mathematical operations, some masking functions and crop and resampling functions. Slicing, projections, mathematical operations, masking, stuff like that is very easy with numpy, so you can easily extend things to what you need. imdata member) such that you can easily work with images in these data formats. The image class is a thin wrapper around typed numpy array objects (the. XDR reading includes NKI compressed images (useful to work with your Elekta images). For the sensitive action recognition issue, the paper provides: (i) a 3 category video database involving non-violent, moderate and extreme violence actions (ii) the conversion of this database into a timed meta-image database from the 2D+Time to 2D conditioning stage described above and (iii) outstanding 2-level and 3-level violence classification results from deep convolutional neural networks operating on meta-image databases.This library supports r/w MetaImage (MHD,ITK), r/w AVSField (.xdr) and read Dicom images. This conversion is such that any 2D frame of the 2D+X data is reshaped as a 1D array indexed by a Hilbert space-filling curve and the third variable X of the initial file format becomes the second variable in the meta-image format. As a consequence of this compressibility, the paper proposes converting the 2D+X data volume into a single meta-image file format, prior to machine learning frameworks. For the data conditioning issue, the paper first highlights that referring 2D spatial convolution to its 1D Hilbert based instance is highly accurate for information compressibility upon tight frames of convolutional networks. The second issue is associated with sensitive action detection in the "2D+Time" case (video clips and image time series). The first issue addressed is conditioning these structured volumes for compatibility with respect to convolutional neural networks operating on image file formats. The paper addresses two issues relative to the machine learning on 2D+X data volumes, where 2D refers to image observation and X denotes a variable that can be associated with time, depth, wavelength, etc.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |