I am trying to find all JPG images within a folder with subfolders that have either width or height below 300px.

This way I want to detect old thumbnails and delete them.

For sure I can find all images using find:

find . -iname "*.jpg" -type f | ...

But what follows after the pipe? Which package can I use to detect picture attributes?


4 Answers 4


You can use identify from imagemagick, and you can use the following command:

find . -iname "*.jpg" -type f -exec identify -format '%w %h %i' '{}' \; | awk '$1<300 || $2<300'

the use of -exec <command> '{}' \; makes sure that your filename can have spaces in them, alternatively you can use

find . -iname "*.jpg" -type f | xargs -I{} identify -format '%w %h %i' {} | awk '$1<300 || $2<300'

where the -I{} takes care of the same thing.

What I like about identify is that you can specify the output format; in this case '%w %h %i' which gives the width, height and full pathname of the image. Then the awk expression only keeps those lines for which the image is smaller than the desired size.

Example of the output:

64 64 ./thumbsup.jpg
100 150 ./photomin.jpg

Edit: If you want the filenames only (for piping to rm for instance), simply change $line in the awk statement to $3, then it will only print the third column.

  • 3
    The awk part can be made shorter: | awk '$1<300||$2<300' or | awk '$1<300||$2<300{print $3}' (when only the 3rd column is needed).
    – alephreish
    Oct 6, 2014 at 18:21
  • @har-wradim Thanks, nice suggestion! Oct 29, 2014 at 13:17
  • 12
    I had to add a newline ('%w %h %i\n') to get it to work for me.
    – qwr
    Jul 24, 2017 at 21:15
  • Sorry for write in a 6 years question, but this is the only result I find on google that partially asnwers my question, basically what I want is find images with or within specific dimensions and then copy them to another directory. I know that cp is used to copy, but I fail at integrate this command with cp.
    – GhostOrder
    Jan 1, 2020 at 19:17
  • 1
    @GhostOrder follow the suggestion of @har-wradim to only print the 3rd column, then add another pipe with something like this | xargs -I {} mv {} /destination/directory/ Jan 2, 2020 at 16:07

This worked for me:

find . -iname "*.png" -type f -exec identify -format '%i %wx%h\n' '{}' \; | grep '75x75'

This is the output sample:

./2520161636481000.png 75x75
./2620160819191100.png 75x75
./2420181545550700.png 75x75

I think the accepted answer is very good, but I wanted to add another possible solution ...

Although I use ImageMagick tools most often now myself, netpbm is an old friend for processing images. You can see the size of any format of image with the command:

anytopnm file | pamfile

This will generate output that looks like:

stdin:  PPM raw, 1650 by 1275  maxval 255

To answer the question of "what follows after the pipe?", I use while read more often than I use xargs because it is more flexible. My netpbm answer to the question looks like this:

find -iname \*.jpg | while read img; do \
  anytopnm "$img" | pamfile | \
    perl -ane 'exit 1 if $F[3]<300 || $F[5]<300' || rm -v "$img"; \

The identify command from the imagemagick package does what you want:

$ identify abc.jpg
abc.jpg JPEG 1952x3264 1952x3264+0+0 8-bit DirectClass 1.111MB 0.000u 0:00.000

Again, you would need to then use grep to sort out the image size.

However, I suspect that unless you have a very wide range of image sizes, it would be easier just to use find to remove JPEG files below a given size:

find -iname '*.jpg' -size -10k -delete

(Worth running without -delete first to check it doesn't find things you want to keep - it won't prompt you before deletion otherwise).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.