I agree with Oliver Krohn - "raw" is what some sensor is delivering without post-processing. Individual to some sensor.

Converting "to raw" would be like trying to restore the raw meat components from some sausage.

Raw is a binary dump of the image (2D) or image volume (3D) with out any header. To convert a single image or single slice in MHA format use VV

https://www.creatis.insa-lyon.fr/rio/vv/

to convert to PNG then use PhotoShop to read PNG and then save a RAW format. Hope this helps, Reuben

@Priyabrata Karmakar : You can not directly read the image directly into Matlab as the .mha format is not supported. 

@Oliver Krohn: I will read the .raw into Matlab. So, firstly, I should convert the .mha to . raw format.

@Reuben Reyes: Thank you Reyes. I converted my all image in MHA format use VV. It's the one that I needed.

Note that you almost always need the header information, you cannot assume that pixel spacing is isotropic or 1mm/pixel but of course separating the file to a text header and a raw file makes loading easier to implement. In MITK and most other medical image computing research software, you can save file as .mhd, which results in two files, a text header file (mhd) and a raw file. You can also save as .nrrd file format and use nrrdread.m and nrrdwrite.m to read/write the files (it supports up to 4D data sets and raw or compressed data) - available at https://app.assembla.com/spaces/slicerrt/subversion/source/HEAD/trunk/MatlabBridge/src/MatlabCommander/commandserver

If you need convenient user interface for importing DICOM data, visualization in 2D/3D, and running your Matlab algorithms on it, then probably the best option is to use 3D Slicer (www.slicer.org) with the MatlabBridge extension (https://www.slicer.org/wiki/Documentation/Nightly/Extensions/MatlabBridge). If you need any help, post a message on http://discourse.slicer.org.

@Andras Lasso: Thank you Lasso. I need medical images for image segmentation. To evaluate my image segmentation algorithm, I need some medical images with their ground truth. So, if I convert images from .mha format to .jpg or .raw formats, I can evaluate my segmentation algorithm in Matlab

Here is an answer someone else posted in ReserchGate which may answer your question: 

Ahmed Elazab · 19.04 · Chinese Academy of Sciences

Hi

You can not directly read the BRATS 2013 or 2014 directly into Matlab as the .mha format is not supported. To tackle this problem, first, I used to convert the .mha to . raw format. Second I read the .raw into Matlab using file manipulation functions (fopen, fread).

To summarize:

1- You can use the MITK visualization software to convert from .mha to .raw

2- Use the following Matlab toolbox to read the .raw

http://www.mathworks.com/matlabcentral/fileexchange/29344-read-medical-data-3d

If you face any problem, you can write it down here.

Good Luck

Jan 7, 2016

@thomas Durand: I used visualization-MITK workbench2016.11.0 software, but it has not such convertor. 

You can use 3D Slicer then. Drag-and-drop to image files to load them. Then in menu click on File / Save scene, choose the file format and click Save.

MATLAB provides tools to read medical data (including .MHA/.MHD), you just need to download (and perhaps purchase) them:

Read Medical Data 3D

version 1.1 (54.5 KB) by Dirk-Jan Kroon

GUI to open most Medical file-formats: Dicom, V3D, ISI, NifTI, Analyze and other formats

https://www.mathworks.com/matlabcentral/fileexchange/29344-read-medical-data-3d?focused=5186778&tab=function&requestedDomain=www.mathworks.com

@Andras Lasso : I also tried with 3D Slicer. It show a message ("Cannot write data file:") when I want to save scene with formats .raw, .bmp,  .jpg, .png and ...

Save as .nrrd and use nrrdread.m available at https://app.assembla.com/spaces/slicerrt/subversion/source/HEAD/trunk/MatlabBridge/src/MatlabCommander/commandserver.

Forget about .bmp, .jpg and other consumer file formats, they are not intended to be used for medical image computing (no standard way of storing essential metadata, compression artifacts, etc).

@Andras Lasso: I read .nrrd file in matlab using  nrrdread.m. But, this function returns only Sagittal slices into "pixelData" Matlab variable . How do we access to Axial slices?