DDBRecon/README at master · particleCT/DDBRecon · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67

#################################################################################
 Distance-driven binning particle CT reconstruction
#################################################################################

This process works for root projection files that have bee preprocessed to produce the 'Full output' from the experimentally acquired data using the pCT collaboration prototype scanner

#################################################################################
1. Running the code
#################################################################################

The process has two steps.

To compile the code, CMakeLists.txt and a compile.sh script are provided in the DDBRecon

The examples assume that the user is working from the 'DDBRecon' Directory

1. run the main script to perform the distance driven binning on each projection

	-'/bin/main projection projection_filter noise_reconstruction CT_filteri [PF_threshold] [prior_file]'

		-projection - the root projection file that is to be used in the reconstruction

		-projection_filter - the type of filtering that should be applied to the projection data (integer 0-3)
			- No filtering (0)
			- dEE & MaxEnergyTransfer & Threshold Filter  (1) **Note this filter is applied to the subsequent filters too**
			- Three-sigma Filter (2)
			- Prior Filter (3)
			*** The filter tags/values must already be in the projection files this code only reads the requested tag from the projections data ***

		-noise_reconstruction (integer 0-1)
			- Image reconstruction (0)
			- Noise reconstruction (1)

		- CT_filter - the filter to be used in the back projection (string)
			- Ram-Lak
			- Cosine

		- PF_threshold *only if using prior filter* - the threshold to use for the prior filter (float 0-1)

		-[prior_file] *optional*  - a root file containing a TH3D object of a previous reconstruction named 'ddb_out'
			- uses a prior reconstruction to detect the hull and perform binning in the following ways:

				- Before the object
					 - particles travel in straight lines following the entrance tracker direction vector (p0)

				- Inside the object
					- Most likely path calculated using cubic splines

				- After exiting the object
					- particles travel in straight lines following the exit tracker direction vector (p1)

	- output (outputs are written to the projection file)
		- TProfile3D 'ddb' - containing the unfiltered result fo the binning
		- Tprofile3D 'ddb_filtered' - filtered 'ddb' orientated at the projection angle


2. run the Reconstruct script on all projection files to create the final reconstruction

	- 'python Reconstruct.py /PATH_TO_PROJECTIONS/'
		-output (outputs written to current directory)
			- recon_out.npz - a numpy .npz array that contains the 3D array of the reconstruction

optional
       - 'python npz-root.py ./recon_out ./output_file_name'
                - creates root file with TH3D of the output array called RSP in the output file specified.
		-NOTE: dimensions are currently hard coded so will need to be changed depending on reconstruction