% Computes motion vectors using *NEW* Three Step Search method
%
% Based on the paper by R. Li, b. Zeng, and M. L. Liou
% IEEE Trans. on Circuits and Systems for Video Technology
% Volume 4, Number 4, August 1994 : Pages 438:442
%
% Input
% imgP : The image for which we want to find motion vectors
% imgI : The reference image
% mbSize : Size of the macroblock
% p : Search parameter (read literature to find what this means)
%
% Ouput
% motionVect : the motion vectors for each integral macroblock in imgP
% NTSScomputations: The average number of points searched for a macroblock
%
% Written by Aroh Barjatya
function [motionVect, NTSScomputations] = motionEstNTSS(imgP, imgI, mbSize, p)
[row col] = size(imgI);
vectors = zeros(2,row*col/mbSize^2);
costs = ones(3, 3) * 65537;
% we now take effectively log to the base 2 of p
% this will give us the number of steps required
L = floor(log10(p+1)/log10(2));
stepMax = 2^(L-1);
computations = 0;
% we start off from the top left of the image
% we will walk in steps of mbSize
% for every marcoblock that we look at we will look for
% a close match p pixels on the left, right, top and bottom of it
mbCount = 1;
for i = 1 : mbSize : row-mbSize+1
for j = 1 : mbSize : col-mbSize+1
% the NEW three step search starts
x = j;
y = i;
% In order to avoid calculating the center point of the search
% again and again we always store the value for it from the
% previous run. For the first iteration we store this value outside
% the for loop, but for subsequent iterations we store the cost at
% the point where we are going to shift our root.
%
% For the NTSS, we find the minimum first in the far away points
% we then find the minimum for the close up points
% we then compare the minimums and which ever is the lowest is where
% we shift our root of search. If the minimum is the center of the
% current window then we stop the search. If its one of the
% immediate close to the center then we will do the second step
% stop. And if its in the far away points, then we go doing about
% the normal TSS approach
%
% more details in the code below or read the paper/literature
costs(2,2) = costFuncMAD(imgP(i:i+mbSize-1,j:j+mbSize-1), ...
imgI(i:i+mbSize-1,j:j+mbSize-1),mbSize);
stepSize = stepMax;
computations = computations + 1;
% This is the calculation of the outer 8 points
% m is row(vertical) index
% n is col(horizontal) index
% this means we are scanning in raster order
for m = -stepSize : stepSize : stepSize
for n = -stepSize : stepSize : stepSize
refBlkVer = y + m; % row/Vert co-ordinate for ref block
refBlkHor = x + n; % col/Horizontal co-ordinate
if ( refBlkVer < 1 || refBlkVer+mbSize-1 > row ...
|| refBlkHor < 1 || refBlkHor+mbSize-1 > col)
continue;
end
costRow = m/stepSize + 2;
costCol = n/stepSize + 2;
if (costRow == 2 && costCol == 2)
continue
end
costs(costRow, costCol ) = costFuncMAD(imgP(i:i+mbSize-1,j:j+mbSize-1), ...
imgI(refBlkVer:refBlkVer+mbSize-1, refBlkHor:refBlkHor+mbSize-1), mbSize);
computations = computations + 1;
end
end
% Now we find the vector where the cost is minimum
% and store it ...
[dx, dy, min1] = minCost(costs); % finds which macroblock in imgI gave us min Cost
% Find the exact co-ordinates of this point
x1 = x + (dx-2)*stepSize;
y1 = y + (dy-2)*stepSize;
% Now find the costs at 8 points right next to the center point
% (x,y) still points to the center
stepSize = 1;
for m = -stepSize : stepSize : stepSize
for n = -stepSize : stepSize : stepSize
refBlkVer = y + m; % row/Vert co-ordinate for ref block
refBlkHor = x + n; % col/Horizontal co-ordinate
if ( refBlkVer < 1 || refBlkVer+mbSize-1 > row ...
|| refBlkHor < 1 || refBlkHor+mbSize-1 > col)
continue;
end
costRow = m/stepSize + 2;
costCol = n/stepSize + 2;
if (costRow == 2 && costCol == 2)
continue
end
costs(costRow, costCol ) = costFuncMAD(imgP(i:i+mbSize-1,j:j+mbSize-1), ...
imgI(refBlkVer:refBlkVer+mbSize-1, refBlkHor:refBlkHor+mbSize-1), mbSize);
computations = computations + 1;
end
end
% now find the minimum amongst this
[dx, dy, min2] = minCost(costs); % finds which macroblock in imgI gave us min Cost
% Find the exact co-ordinates of this point
x2 = x + (dx-2)*stepSize;
y2 = y + (dy-2)*stepSize;
% the only place x1 == x2 and y1 == y2 will take place will be the
% center of the search region
if (x1 == x2 && y1 == y2)
% then x and y still remain pointing to j and i;
NTSSFlag = -1; % this flag will take us out of any more computations
elseif (min2 <= min1)
x = x2;
y = y2;
NTSSFlag = 1; % this flag signifies we are going to go into NTSS mode
else
x = x1;
y = y1;
NTSSFlag = 0; % This value of flag says, we go into normal TSS
end
if (NTSSFlag == 1)
% Now in order to make sure that we dont calcylate the same
% points again which were in the initial center window we take
% care as follows
costs = ones(3,3) * 65537;
costs(2,2) = min2;
stepSize = 1;
for m = -stepSize : stepSize : stepSize
for n = -stepSize : stepSize : stepSize
refBlkVer = y + m; % row/Vert co-ordinate for ref block
refBlkHor = x + n; % col/Horizontal co-ordinate
if ( refBlkVer < 1 || refBlkVer+mbSize-1 > row ...
|| refBlkHor < 1 || refBlkHor+mbSize-1 > col)
continue;
end
if ( (refBlkVer >= i - 1 && refBlkVer <= i + 1) ...
&& (refBlkHor >= j - 1 && refBlkHor <= j + 1) )
continue;
end
costRow = m/stepSize + 2;
costCol = n/stepSize + 2;
if (costRow == 2 && costCol == 2)
continue
end
costs(costRow, costCol ) = costFuncMAD(imgP(i:i+mbSize-1,j:j+mbSize-1), ...
imgI(refBlkVer:refBlkVer+mbSize-1, refBlkHor:refBlkHor+mbSize-1), mbSize);
computations = computations + 1;
end
end
% now find the minimum amongst this
[dx, dy, min2] = minCost(costs); % finds which macroblock in imgI gave us min Cost
% Find the exact co-ordinates of this point and stop
x = x + (dx-2)*stepSize;
y = y + (dy-2)*stepSize;
elseif (NTSSFlag == 0)
% this is when we are going about doing normal TSS business
costs = ones(3,3) * 65537;
costs(2,2) = min1;
stepSize = stepMax / 2;
while(stepSize >= 1)
for m = -stepSize : stepSize : stepSize
for n = -stepSize : stepSize : stepSize
refBlkVer = y + m; % row/Vert co-ordinate for ref block
refBlkHor = x + n; % col/Horizontal co-ordinate
if ( refBlkVer < 1 || refBlkVer+mbSize-1 > row ...
|| refBlkHor < 1 || refBlkHor+mbSize-1 > col)
continue;
end
costRow = m/stepSize + 2;
costCol = n/stepSize + 2;
if (costRow == 2 && costCol == 2)
continue
end
costs(costRow, costCol ) = costFuncMAD(imgP(i:i+mbSize-1,j:j+mbSize-1), ...
imgI(refBlkVer:refBlkVer+mbSize-1, refBlkHor:refBlkHor+mbSize-1), mbSize);
computations = computations + 1;
end
end
% Now we find the vector where the cost is minimum
% and store it ... this is what will be passed back.
[dx, dy, min] = minCost(costs); % finds which macroblock in imgI gave us min Cost
% shift the root for search window to new minima point
x = x + (dx-2)*stepSize;
y = y + (dy-2)*stepSize;
stepSize = stepSize / 2;
costs(2,2) = costs(dy,dx);
end
end
vectors(1,mbCount) = y - i; % row co-ordinate for the vector
vectors(2,mbCount) = x - j; % col co-ordinate for the vector
mbCount = mbCount + 1;
costs = ones(3,3) * 65537;
end
end
motionVect = vectors;
NTSScomputations = computations/(mbCount - 1);
