Vtyjkyuk

I want to highlight some part of an image and hiding the remaining part. My minimum working example is below. Any help will be highly appreciated. Thanks

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

begindocument

beginframe

begintikzpicture
node[anchor = south west, inner sep = 0] (A) at (0, 0) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio]Effile.jpeg;

draw[red, ultra thick] (1.05, 5.1) rectangle (2.49, 4.3);

fill [draw = none, fill = white, fill opacity = 1.0] (1.05, 5.1) -- (2.49, 5.1) -- (2.49, 1.05) -- cycle;

endtikzpicture

endframe

enddocument

Edit
With the overlay-beamer-styles library this can be done quite elegantly.
Apparently one does not need the overlay-beamer-styles library, because the Tikz macros work with the overlay syntax of beamer!

I include the image in the same way, but name the node image. Also, it doesn't matter where the node is placed.

begintikzpicture
 node[anchor = south west, inner sep = 0] (image) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

Then I can define a scope that has (0,0) in the bottom left corner of image, (1,0) in the bottom right corner, (0,1) in the top left corner and (1,1) in the top right corner.

 beginscope[shift=(image.south west),x=(image.south east),y=(image.north west)]

Drawing a grid that scales with the image is now straightforward. This makes it easy to recognize the place of the paths needed in the next step.

 % Draw a overlaying grid to easily see where the rectangles must be, this can be commented out in the final version
 draw foreach xy in 0,0.1,...,1.001
 (xy,0) -- node[pos=0,below]pgfmathprintnumberxy (xy,1)
 (0,xy) -- node[pos=0,left]pgfmathprintnumberxy (1,xy);

Having identified the paths of the segment I want to highlight, I can add these to my foreach loop, which draws a big white rectangle on top of the image, excludes the path I specified, and draws an additional red rectangle around the highlighted segment. Additionally, the overlay-beamer-styles library provides the visible on=<> key, which can be used with the counter in the foreach.

 foreach[count=i] mypath in 
 (0.1,0.55) rectangle (0.25,0.65),
 (0.45,0.85) rectangle (0.6,0.95),
 (0.5,0.5) circle (1cm),
 (0.75,0.3) rectangle (0.85,0.35) (0.25,0.75) circle (1.5cm)
 
 filldraw<i>[white,opacity=0.75,even odd rule] (0,0) rectangle (1,1) mypath;
 draw<i>[red,ultra thick] mypath;
 

Finally close the scope and the tikzpicture.

 endscope
endtikzpicture

This results in:

where the grid is still visible, and the white rectangle is 'see-through', but for a final version this can be made opaque and the grid can be hidden.

Total MWE (also note that I draw the grid on top of the images and white rectangle, because otherwise it would not be very helpful):

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

begindocument
 beginframe
 begintikzpicture
 node[anchor = south west, inner sep = 0] (image) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

 beginscope[shift=(image.south west),x=(image.south east),y=(image.north west)]

 foreach[count=i] mypath in 
 (0.1,0.55) rectangle (0.25,0.65),
 (0.45,0.85) rectangle (0.6,0.95),
 (0.5,0.5) circle (1cm),
 (0.75,0.3) rectangle (0.85,0.35) (0.25,0.75) circle (1.5cm)
 
 filldraw<i>[white,even odd rule] (0,0) rectangle (1,1) mypath;
 draw<i>[red,ultra thick] mypath;
 

% Draw a overlaying grid to easily see where the rectangles must be, this can be commented out in the final version
% draw foreach xy in 0,0.1,...,1.001
% (xy,0) -- node[pos=0,below]pgfmathprintnumberxy (xy,1)
% (0,xy) -- node[pos=0,left]pgfmathprintnumberxy (1,xy);
 endscope

 endtikzpicture
 endframe
enddocument

Edit 2

Due to the fancy example shown by @marmot in his answer, I had to add something fancy myself :)

This can be done with:

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

begindocument
 beginframe
 begintikzpicture
 node[anchor = south west, inner sep = 0] (image) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

 beginscope[shift=(image.south west),x=(image.south east),y=(image.north west)]
 clip (0,0) rectangle (1,1);
 foreach[count=i] myDim in 0,5,...,106
 beginscope[shift=(0.5,0.5)]
 filldraw<i>[white,opacity=1,even odd rule,] (-0.5,-0.5) rectangle (0.5,0.5) (0,1.38*myDim mm)
 -- ++(288:myDim mm) -- ++( 0:myDim mm)
 -- ++(216:myDim mm) -- ++(288:myDim mm)
 -- ++(144:myDim mm) -- ++(216:myDim mm)
 -- ++( 72:myDim mm) -- ++(144:myDim mm)
 -- ++( 0:myDim mm) -- cycle;
 draw<i>[red,ultra thick] (0,1.38*myDim mm)
 -- ++(288:myDim mm) -- ++( 0:myDim mm)
 -- ++(216:myDim mm) -- ++(288:myDim mm)
 -- ++(144:myDim mm) -- ++(216:myDim mm)
 -- ++( 72:myDim mm) -- ++(144:myDim mm)
 -- ++( 0:myDim mm) -- cycle;
 endscope
 
 endscope

 endtikzpicture
 endframe
enddocument

Edit 3

This will (probably _^(maybe)) be the last one:

Done with:

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

usetikzlibraryshapes

begindocument
 beginframe
 begintikzpicture
 node[anchor = south west, inner sep = 0,outer sep=0.5pgflinewidth] (image) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

 clip (image.south west) rectangle (image.north east);

 foreach[count=k] myDim in 0,0.2,...,3.4
 ifnumk>1
 defmypath
 [scale=myDim^2] (image.center)
 foreach myPhi [count=i,evaluate=i as j using 0.5*mod(i,2)+0.5] in 0,36,...,359
 ifnumi=1
 +(90+50*sqrt(k)+myPhi:j)
 else
 -- +(90+50*sqrt(k)+myPhi:j)
 fi
 -- cycle
 
 else
 defmypath
 fi

 filldraw<k>[white]
 (image.south west) rectangle (image.north east) mypath;
 draw<k>[red,ultra thick] mypath;
 

 endtikzpicture
 endframe
enddocument

Edit 4

This really is the final one:

Done with the clip command (sneakingly stolen from @marmot's answer). Also I defined a use node cs style which takes a node name as argument and makes it easier to draw in that relative coordinate system:

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

usetikzlibrarycalc

tikzset
 use node cs/.style=shift=(#1.south west),x=(#1.south east),y=(#1.north west)


begindocument
 beginframe
 begintikzpicture
 node[anchor = south west, inner sep = 0,opacity=0] (image) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

 clip (image.south west) rectangle (image.north east);

 foreach myDim [count=i] in0,...,30
 clip<i>[use node cs=image] let 
 n1=max(myDim,0),
 n2=max(myDim - 1,0),
 n3=max(myDim - 2,0),
 n4=max(myDim - 3,0),
 n5=max(myDim - 4,0)
 in 
 (0.1,0.2) circle (n1 mm)
 (0.6,0.4) circle (n2 mm)
 (0.8,0.2) circle (n5 mm)
 (0.5,0.3) circle (n1 mm)
 (0.1,0.8) circle (n5 mm)
 (0.4,0.9) circle (n3 mm)
 (0.7,0.7) circle (n1 mm)
 (0.2,0.5) circle (n3 mm)
 (0.9,0.6) circle (n2 mm)
 (0.9,0.8) circle (n1 mm);
 
 node[anchor = south west, inner sep = 0] includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio] Effile.jpeg;

 endtikzpicture
 endframe
enddocument

I think Max gave you already a perfect answer. This here is just for fun and to show a slightly different approach using clip. You could just clip the desired regions using only.

documentclassbeamer
usepackagetikz

begindocument

beginframe

begintikzpicture
coordinate (O) at (0,0);
node[anchor = south west, inner sep = 0,opacity=0] (A) at (O) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio]Effile.jpeg;
only<1>clip (A.center) circle (1cm);
only<2>clip (A.center) circle (2cm);
only<3>clip (1.05, 5.1) rectangle (2.49, 4.3);
only<4>
node[anchor = south west, inner sep = 0] at (O) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio]Effile.jpeg;
only<3>draw[red, ultra thick] (1.05, 5.1) rectangle (2.49, 4.3);
endtikzpicture
endframe
enddocument

And one can of course get rid of only and do other things like this.

documentclassbeamer
usepackagetikz

begindocument

beginframe

begintikzpicture
coordinate (O) at (0,0);
node[anchor = south west, inner sep = 0,opacity=0] (A) at (O) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio]Effile.jpeg;
foreach X [count=Y] in 0.9,0.8,...,0.1
 
 path (A.south west) -- (A.north east) coordinate[pos=X*0.5] (BL)
 coordinate[pos=1-0.5*X] (TR);
 clip<Y> (BL) rectangle (TR);
 
node[anchor = south west, inner sep = 0] at (O) includegraphics[width = 0.8paperwidth, height = 10cm, keepaspectratio]Effile.jpeg;
endtikzpicture
endframe
enddocument

EDIT: I hadn't seen Max' super nice updates. So I had to rewrite his nice star code in a somewhat shorter way. One can load shapes.geometric, which also has a star (and many other shapes), and use path picture to clip the pic against that shape. ;-)

documentclassbeamer
usepackagetikz
usetikzlibraryshapes.geometric
begindocument
 beginframe
 begintikzpicture
 path[use as bounding box] (-6,-4) rectangle (6,4);
 foreach X [count=Y]in 2,2.4,...,16
 node<Y> [rotate=X*60,star, star point height=0.25*X cm, minimum size=X cm, draw,
 path picture=
 nodeincludegraphics[width = 0.8paperwidth, height = 10cm,
 keepaspectratio] Effile.jpeg;]
 (image) ;
 endtikzpicture
 endframe
enddocument

May be this could help. By properly naming the node you can use a fill=<color> with fill opacity=<num>. By naming the node we can easily access the coordinates of the vertices, thereby, we can flow-through the region we want:

documentclassbeamer
usepackage[T1]fontenc
usepackagetikz

begindocument

beginframe
% yes this is the original pic just for comparing
begintikzpicture
node[anchor = south west, inner sep = 0] (A) at (0, 0) includegraphics[width = 0.2paperwidth, height = 10cm, keepaspectratio]example-image-a;

endtikzpicture
% the actual stuff, the highlighted picture!
begintikzpicture
node[anchor = south west, inner sep = 0] (A) at (0, 0) includegraphics[width = 0.2paperwidth, height = 10cm, keepaspectratio]example-image-a;

draw[red,ultra thick] (0,0) rectangle (0.5,0.5);
fill [draw=none, fill=white, fill opacity=1] 
(0,0.5) -- 
(A.north west) -- (A.north east) -- (A.south east) -- 
(0.5,0) -- (0.5,0.5) -- cycle;

endtikzpicture

endframe

enddocument

which gives you: