1

I'm creating an open Neural Network that could identify if the photo is blurred. This way I need to find many photos of objects at close range, and the focus should be upset.

The same as you can see in the two following photos: image description image description

What are these photos called? What can I push to the search board and get the same photos?

I already tried to use the "Bokeh Effect" search query but it gives photos with unwanted effects:

  1. It usually has focus on at least one object;
  2. It usually has blurred light and lamps:

image description

How can I find the same photos (as I have) without any focus and lights?

  • 1
    Good luck with that, typically people don't post bad pictures. Or google for "Sorry for the picture quality", but your neural net will be mostly learning with pictures of pets... – xenoid Mar 6 at 13:00
  • I'm not familiar with neural nets, but can't you take sharp photos, apply a high quality blur on them to simulate them being out of focus and then train your neural net on that? – Saaru Lindestøkke Mar 6 at 13:10
  • Re, "bokeh," You are exactly right. The literal meaning of the Japanese word is "blurry," but people usually only say "bokeh" when small, out-of-the-plane-of-focus highlights in the subject (e.g., distant, bright lights in an otherwise dark background) show up as bright silhouettes of the lens aperture (usually circles, or polygons) in the photograph. – Solomon Slow Mar 6 at 15:10
  • 1
    One potential issue is that often a photo one would call out of focus isn't out of focus across the whole image (as in your examples). A very common issue is where the intended subject is not in the zone of acceptable focus, but other parts of the foreground or background are. Of course here I'm assuming that you will be looking at photographs taken by and for humans. – David Rouse Mar 6 at 18:59
  • Hi Egor, Welcome to Photography. Another StackExchange site may be better suited to help you find a better word for your image edge-quality tag name. Try english.stackexchange.com for collaboration on other terms such as "blurred, not sharp, out-of-focus, soft, or even fuzzy" images. – Stan Mar 7 at 14:54
4

IMO, a neural net would be overkill for identifying blurred images. Just run an edge detection filter on it. If there are no strong edges anywhere in the frame, then it's blurred.

| improve this answer | |
2

Bad photos (no focus). What are these photos called?

I am highly qualified to pontificate on this subject.( not that i always feel the need to be qualified on a subject in order to pontificate on it. )

I have made many many of these photos over the years.

I call them "mistakes".

Sometimes these mistakes even cause me to question if i should even own a camera or am if i am experiencing a loss of vision. ( Physical vision as well as metaphorical vision )

In fact these mistakes seem to be inherent in my work, They pop up frequently as if i had never learned how to focus a camera.

I generally do not keep these photos and i have never had the need to catalogue them or have software to aggregate them into one place. ( digital files that is, i have binders full of analogue mistakes but they are all mixed in with properly focused negatives and even some of those mistakes are Happy Mistakes )

I love happy mistakes, even a blind squirrel finds a nut sometimes.

When transferring files to my computer I use my eyes, flawed as they are, and Adobe Bridge to view my digital files on a monitor and determine which files are mistakes or otherwise not worth keeping and then delete them.

Many times viewing the files at monitor size shows that files i thought were good on the cameras LCD screen are in fact mistakes.

| improve this answer | |
1

I would search for "out of focus photography" or "blurred photo".

In digital image processing software "unsharp masking (USM)" is used as an image sharpening technique. Maybe you'll find some useful images under this term too.

Have a nice day.

| improve this answer | |
1

"Defocused" is the term I'd look for. As opposed to "unfocused", its rare to use it out of optical contexts.

| improve this answer | |

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.