Upload
amit-kumar
View
217
Download
0
Embed Size (px)
Citation preview
8/7/2019 Scratch Detection and Removal From Old Videos
1/19
8/7/2019 Scratch Detection and Removal From Old Videos
2/19
Scratch Detection And Restoration
Why Important?
Photographic archives are being
transferred to digital form
Involves scanning of transmissive and
reflective media, namely slides or
negatives, and prints.
Inevitably introduces new artifacts such
as dust particles and scratches which
causes degradation of quality
8/7/2019 Scratch Detection and Removal From Old Videos
3/19
Why Automate?
Because of the huge number of the involved frames
Detection and Removal
The problem of automatic defect removal
is divided into two stages: detection and
removal
Detection: Defective piels identified
!econstruction: Defective piels are
replaced with estimated original piels
8/7/2019 Scratch Detection and Removal From Old Videos
4/19
Hurdles?
Inaccurate detection :"eparating true defects and
intrinsic image features, such as
edges, line segments and various
tetures.
Inaccurate detection leads to
distortion of intrinsic image
features during the reconstruction
stage
8/7/2019 Scratch Detection and Removal From Old Videos
5/19
Artiacts in Old Films
The main defect in old films are dust spots, line
scratches, abrasion, blotch, film unsteadiness
and intensity flic#er
$ine scratches are one of the most annoying
artifacts in old films
%aused by particle caught in film transport
mechanism, and when film passes over the
particles material being abraded and line
scratches are formed.
8/7/2019 Scratch Detection and Removal From Old Videos
6/19
&ll about "cratch
The width of line scratches are between '()
piels
In the center of line scratches, the intensity of
the piel shows a very little fluctuation
The level of the intensity values for line
scratches is higher than that of image edges
Typically they are vertical *angle to the verticalupto + degrees
$oss of color information in the region is
complete
*usually have bright or dar# color!
8/7/2019 Scratch Detection and Removal From Old Videos
7/19
Scratch Detection
Two ma-or types: i"patial methods
ii Temporal methods
"patial( "patial detection methods usesstrong vertical scratch property
Temporal( Trac#ing of scratch in motion
compensated image frames
8/7/2019 Scratch Detection and Removal From Old Videos
8/19
OW"#$ased line scratch detection
/0 is an undecimated wavelet transform withtranslation invariant
It locates the position of line scratches in each
scale as same as the position in the spatial domain
without any information damaged during thewavelet transform
1rom this point of view, /0 is more appropriate
than decimated wavelet transforms in image post(
processing
&llows us to design a more efficacious and
precise position
for scratch than traditional methods
8/7/2019 Scratch Detection and Removal From Old Videos
9/19
%roposed ethod
'enerali(ed Wavelet ")pansion $asedAlgorithm
To locate the line artifact and represent them
precisely, the algorithm adopts ver ( complete
/avelet 0pansion
*/0 representation of original image
&t the edge of line scratches area, hori2ontal
intensity
variation is much larger than vertical intensity
variation
8/7/2019 Scratch Detection and Removal From Old Videos
10/19
Flo*chart o Algorithm
Grey ScaleImage
ColoredImage
Im2doubleDouble values
OWE(A,H,V,D)
Sum V(Vertcal)
!ocal V"a#ma ($)
$() % $(&)'Scratc
V must ave at least *+ + as above
average
rgb2gray
"as- (tose .s)
8/7/2019 Scratch Detection and Removal From Old Videos
11/19
Steps
%onvert the image to grey scale if it is !3B andchange the values in double to apply /avelet
transform in net step.
Ta#e ver ( complete wavelet epansion */0.
&ssume that the original image has m by n piels,
and denote / v*,y as first scale vertical wavelet
decomposition coefficient at the position of *, y,
then the sum is: nV M W v (x) = W v ( x , y ), ----- (1)
y=1
8/7/2019 Scratch Detection and Removal From Old Videos
12/19
/e find the local maima and minima and chec# if
there is local maima and minima in nearby piel
based on scratch width*+ piel both side.
Based on the fact that there are sharp hori2ontal
variations at both sides of line scratches and local
maimum and local minimum will be found at the left
side and right side of linescratches respectively
xW Max={x|V MWv(x)>V MWv(x-1)&V MWv(x)>V
MWv(x+1),xZ,1
8/7/2019 Scratch Detection and Removal From Old Videos
13/19
1or actual position of scratches
& 456+ bloc# is ta#en as local bloc# image andpiel values are compared with average.
If 75 piel in a column has more or less than
average value than this column has a scratch in
particular bloc#.
8/7/2019 Scratch Detection and Removal From Old Videos
14/19
Scratched Image Scratch Detected
8/7/2019 Scratch Detection and Removal From Old Videos
15/19
!estoration
!estoration improves the sub-ective visual
quality of archive
To restore the film without causing distortion to
areas of the frames that are not affected, the
location of scratch must be located precisely
8oreover a fully automatic restoration system is
very hard to develop, because the correction of
misalignment of brightness, detection and removal
of shot and scratch, noise reduction, etc.
8/7/2019 Scratch Detection and Removal From Old Videos
16/19
&ssumption: !egions affected by scratches do
not contain original information
"olution: Propagate neighboring clean
information into the degraded area
The neighboring information can be found in thesame frame or also in the preceding and
successive frame eploiting the temporal
coherency
8/7/2019 Scratch Detection and Removal From Old Videos
17/19
Functional $loc+
diagram
Dgtal
DegradedSe/ue0ce1mage
A0alogue
DegradedSe/ue0ce1mage
ScratcDetector OWE mas-
Scratces
(!esse0 alseDetecto0)
Com3ar0g
rames(O0ly or vdeos)I0ter3olato0
$estored
Se/1mage
Dgtal E0coder
8/7/2019 Scratch Detection and Removal From Old Videos
18/19
!esult: Images scratch
removed
Scratced Image Scratc$emoved
8/7/2019 Scratch Detection and Removal From Old Videos
19/19
,HA-. /O0