weekly meeting - miscellaneous updates · weekly meeting miscellaneous updates j er emie decock cea...
Post on 20-Aug-2020
9 Views
Preview:
TRANSCRIPT
Benchmark Energy conservation Analysis Conclusion References
Weekly meetingMiscellaneous updates
Jeremie Decock
CEA Saclay - Irfu/SAp
September 28, 2016
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Benchmark
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Definition
Definition
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Definition
Benchmark function
The error function E is given by:
E(s, s∗) = (E1(s, s∗), E2(s, s∗))T
E1(s, s∗) = mean
(abs
(s∑i si− s∗∑
i s∗i
))E2(s, s∗) =
abs (∑
i si −∑
i s∗i )∑
i s∗i
Where:
I s is the output image (the ”cleaned” image) ∈ IRd
I s∗ is the reference image (the ”clean” image) ∈ IRd
I i is the index of a PMT (i.e. of a pixel) within an image
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Gammas
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
I ASTRI mini-array (calibrated data)
I Telescopes 1 to 33 only (ASTRI)
I Polychromatic event set
I Input files: sapcta:/dsm/manip/cta/DATA/astri mini array/fits/gamma/
I Source file:sapcta:/dsm/manip/cta/DATA/astri mini array/gamma/run1001.simtel.gz
I Num samples: 14899 images
Cleaning algorithms:
I Tailcut: JD’s implementation
I FFT: Numpy implementation
I Wavelets: Cosmostat Sparce2D (mr transform) b-Spline wavelet transform
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
FFT Tailcut WT10-5
10-4
10-3
10-2
Sco
reE1 (γ)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-4 10-3
Score
100
101
102
103
104C
ount
E1 (γ)
FFT
Tailcut
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
13.5
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
80
160
240
320
400
480
560
0
80
160
240
320
400
480
560
80
160
240
320
400
480
560
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
9.0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
2
4
6
8
10
12
14
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
13.5
15.0
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
13.5
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0.0
1.5
3.0
4.5
6.0
7.5
0.0
0.6
1.2
1.8
2.4
3.0
3.6
4.2
4.8
5.4
6.0
2.4
3.0
3.6
4.2
4.8
5.4
6.0
6.6
7.2
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
150
300
450
600
750
900
1050
1200
0
150
300
450
600
750
900
1050
1200
150
300
450
600
750
900
1050
1200
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
FFT Tailcut WT10-4
10-3
10-2
10-1
100
101
102
103
Sco
reE2 (γ)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-4 10-3 10-2 10-1 100 101 102 103
Score
100
101
102
103
104
Count
E2 (γ)
FFT
Tailcut
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
9.0
0.0
0.6
1.2
1.8
2.4
3.0
3.6
4.2
4.8
5.4
6.0
1.5
3.0
4.5
6.0
7.5
9.0
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
250
500
750
1000
1250
1500
1750
2000
0
500
1000
1500
2000
2500
3000
3500
4000
250
500
750
1000
1250
1500
1750
2000
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
0.0
0.8
1.6
2.4
3.2
4.0
4.8
5.6
6.4
0
1
2
3
4
5
6
7
8
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Bad WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
9.0
0.0
0.8
1.6
2.4
3.2
4.0
4.8
5.6
6.4
7.2
8.0
2.4
3.2
4.0
4.8
5.6
6.4
7.2
8.0
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
Good WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
250
500
750
1000
1250
1500
1750
2000
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
250
500
750
1000
1250
1500
1750
2000
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-5 10-4 10-3 10-2
Score 0
10-4
10-3
10-2
10-1
100
101
102
103
Sco
re 1
Null scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-5 10-4 10-3 10-2
Score 0
10-4
10-3
10-2
10-1
100
101
102
103
Sco
re 1
FFT scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-4 10-3 10-2
Score 0
10-4
10-3
10-2
10-1
100
101
102
Sco
re 1
Tailcut scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gammas
10-4 10-3 10-2
Score 0
10-2
10-1
100
101
102
103
Sco
re 1
WT scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Protons
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
I ASTRI mini-array (calibrated data)
I Telescopes 1 to 33 only (ASTRI)
I Polychromatic event set
I Input files: sapcta:/dsm/manip/cta/DATA/astri mini array/fits/proton/
I Source file:sapcta:/dsm/manip/cta/DATA/astri mini array/proton/run10001.simtel.gz
I Num samples: 2203 images
Cleaning algorithms:
I Tailcut: JD’s implementation
I FFT: Numpy implementation
I Wavelets: Cosmostat Sparce2D (mr transform) b-Spline wavelet transform
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
FFT Tailcut WT10-5
10-4
10-3
10-2
Sco
reE1 (protons)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-4 10-3
Score
100
101
102
103C
ount
E1 (γ)
FFT
Tailcut
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
FFT Tailcut WT10-5
10-4
10-3
10-2
10-1
100
101
102
103
104
Sco
reE2 (protons)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-5 10-4 10-3 10-2 10-1 100 101 102 103 104
Score
100
101
102
103
CountE2 (protons)
FFT
Tailcut
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Bad FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1
2
3
4
5
6
7
8
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Good FFT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
50
100
150
200
250
300
350
0
40
80
120
160
200
240
280
320
360
40
80
120
160
200
240
280
320
360
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Bad Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
2
4
6
8
10
12
0.0
0.6
1.2
1.8
2.4
3.0
3.6
4.2
4.8
5.4
6.0
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Good Tailcut example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
2
4
6
8
10
12
14
0.0
1.5
3.0
4.5
6.0
7.5
9.0
10.5
12.0
13.5
15.0
0
2
4
6
8
10
12
14
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Bad WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
1.5
0.0
1.5
3.0
4.5
6.0
7.5
0.0
0.4
0.8
1.2
1.6
2.0
2.4
2.8
3.2
3.6
4.0
2.4
3.0
3.6
4.2
4.8
5.4
6.0
6.6
7.2
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
Good WT example
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0
250
500
750
1000
1250
1500
1750
2000
0
400
800
1200
1600
2000
2400
2800
3200
250
500
750
1000
1250
1500
1750
2000
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-5 10-4 10-3 10-2
Score 0
10-4
10-3
10-2
10-1
100
101
102
103
104
Sco
re 1
Null scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-5 10-4 10-3 10-2
Score 0
10-3
10-2
10-1
100
101
102
103
104
Sco
re 1
FFT scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-4 10-3 10-2
Score 0
10-5
10-4
10-3
10-2
10-1
100
101
102
103
Sco
re 1
Tailcut scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Protons
10-4 10-3 10-2
Score 0
10-1
100
101
102
103
104
Sco
re 1
WT scores correlation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Energy conservation
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Definition
Definition
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Definition
Definition
The error on energy is given by:∑i si −
∑i s
∗i∑
i s∗i
Where:
I s is the output image (the ”cleaned” image) ∈ IRd
I s∗ is the reference image (the ”clean” image) ∈ IRd
I i is the index of a PMT (i.e. of a pixel) within an image
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gamma
Gamma
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gamma
0.2 0.0 0.2 0.4 0.6 0.8 1.0
Score 1e3
100
101
102
103
104
Count
FFT
FFT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gamma
1 0 1 2 3 4 5 6
Score 1e1
100
101
102
103
104
105
Count
Tailcut
Tailcut
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Gamma
0.2 0.0 0.2 0.4 0.6 0.8 1.0
Score 1e3
100
101
102
103
104
Count
WT
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Proton
Proton
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Proton
1 0 1 2 3 4 5
Score 1e3
100
101
102
103
104
Count
FFT
FFT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Proton
0.5 0.0 0.5 1.0 1.5 2.0 2.5
Score 1e2
100
101
102
103
104
Count
Tailcut
Tailcut
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Proton
0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Score 1e3
100
101
102
103
104
Count
WT
WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Analysis
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Background > 0 with FFT and WT
Background > 0 with FFT and WT
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Background > 0 with FFT and WT
Look at the scale of the cleaned image...
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
0
2
4
6
8
10
12
14
16
18
20
4
6
8
10
12
14
16
18
20
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Background > 0 with FFT and WT
It’s more obvious with ’vmin=0’ in matplotlib...
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Input image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Reference image
0 5 10 15 20 25 30 350
5
10
15
20
25
30
35
Cleaned image
0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
0
2
4
6
8
10
12
14
16
18
20
0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Background > 0 with FFT and WT
I It explains the shift on all plots for FFT and WT (in additionto the wrong threshold ?)
I It should explain why the classification works despite thatmistake (the background is ignored by the classifier)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
Conclusion
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
TODO
TODO
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
TODO
TODO
I fix the issue with the non null background with FFT and WTby using the right option in mrtransform / mrfilter (alsoremember that some pixels in the input image have negativevalue...)
I use vmin=0 in matplotlib to make this mistake more visible
I plot score vs Npe , ...
I add metadata into input FITS files (EMC , ...)
I use these metadata to plot log(EMC ) vs Npe , score vslog(EMC ), ...
I plot the Chi2
I tag images where the signal is on the border
I optimize meta parameters (multicriteria optimization)
Decock CEA Saclay - Irfu/SAp
Weekly meeting
Benchmark Energy conservation Analysis Conclusion References
References I
Decock CEA Saclay - Irfu/SAp
Weekly meeting
top related