SLIDE 1
Event-based color camera
Alexandre Marcireau
1
2
Event-based color camera Alexandre Marcireau 2019-03-04 PhD - - PowerPoint PPT Presentation
Event-based color camera Alexandre Marcireau 2019-03-04 PhD - - PowerPoint PPT Presentation
Event-based color camera Alexandre Marcireau 2019-03-04 PhD advisor: Pr. R. Benosman 1 Frames and events t t y y x x synchronous, global shutter asynchronous, independent pixels 2 ATIS (Asynchronous time-based image sensor) DVS
SLIDE 2
SLIDE 3
3
ATIS (Asynchronous time-based image sensor)
DVS + event-based absolute exposure measurements I ∝ 1 / Δt
SLIDE 4
4
Three-chip event-based camera
Color sensor
SLIDE 5
5
ATIS ATIS ATIS
Generating color events
SLIDE 6
6
Task: find the mean position of wooden pieces over time, using only color information
Validation: color-based tracking
SLIDE 7
7
red green
RGB, red + green + blue = 1
1 1 a b c a b c a* b*
CIEL*a*b*, L = 50
Perceptually uniform with respect to the Euclidean distance
Using CIEL*a*b* to approximate human perception
Color spaces
SLIDE 8
8
Color calibration
Macbeth color checker Expected values as a function of the camera measurements. The red line is a mean-square errors fit. The green line is a non-linear optimisation fit with CIEL*a*b* distances as heuristic.
SLIDE 9
9
Color calibration - results
L* b* a* L* b* a*
Measured points in CIEL*a*b* space Mean distance in CIEL*a*b* between measurements and expected values, expressed as a fraction the largest distance between expected values.
SLIDE 10
10
Example
- ptimized
linear calibration channel-wise tone mapping
SLIDE 11
11
Color signatures
Bivariate gaussian distributions are used to describe the color measurements
SLIDE 12
12
Moving mean example with λ = 0.25
λ ∈ [0, 1] is the inertia parameter mi = λ mi-1 + (1 - λ) ei
Event-based moving mean
SLIDE 13
13
find the closest signature merge the event streams update the
- bject’s
mean position
ATIS ATIS
convert the color to CIEL*a*b*
ATIS
Event-based pipeline
SLIDE 14
14
Results - controlled scene
SLIDE 15
15
Outdoor scenes
fixed camera, moving objects fixed objects, moving camera
SLIDE 16
16
Time x y Time Time x y Time
Results - outdoor scenes
x (pixels) 50 1 2 5 3 4 250 Time (s) 150 Time (s) 140 120 100 y (pixels) 1 2 5 3 4 x (pixels) 50 3.0 4.5 3.5 4.0 200 150 Time (s) 100 3.0 4.5 3.5 4.0 Time (s) 120 100 80 y (pixels)
SLIDE 17
17
Discussion and future work
continuous spatial analog signal pixels / photoreceptor cells discrete spatial analog signal analog processing analog-to-digital conversion discrete spatial digital signal digital processing combine color channels? mechanical feedback 1 2 3 4
- r
SLIDE 18
18