How do I enhance low quality JPEG images?

AI Thread Summary
Enhancing low-quality and low-resolution JPEG images can be challenging due to the need for algorithms that add data not originally present. While Photoshop offers advanced tools for this purpose, free alternatives exist. Programs like Paint.NET feature an "unsharp mask" plugin that sharpens edges but can also increase noise, making images grainy. Users are advised to apply this effect cautiously for best results. Other options include online tools like Topaz Labs' DeJPEG, which blurs JPEG artifacts, and GIMP with the wavelet denoise plugin for noise reduction. However, it's important to note that while these enhancements can improve appearance, they do not add actual information to the image, and excessive processing may introduce artifacts that degrade quality. For specific issues like blurred text, tools like SmartDeblur can be effective, but results may vary based on the original image quality and compression levels. Ultimately, understanding the limitations of these tools is crucial for managing expectations regarding image enhancement.
ElliotSmith
Messages
167
Reaction score
104
Is there any way to enhance low quality and low resolution JPEG images? Is there a program I can use or something?

I have dozens of grainy and poor quality JPEGs that I would like to sharpen up, but I do not own photoshop.
 
Computer science news on Phys.org
Enhancing low quality photos is tough because you need some algorithm to add in data to the image that didn't exist. Photoshop does have some advanced tools for this. Free options do exist but I'm not sure how much they will help. Certainly sharpen filters are common though.

Try http://www.getpaint.net
 
To sharpen a picture, look at the "unsharp mask" plugin for paint.net. It's name is deceptive. It actually sharpens edges. It must be used judiciously because it is easy to overdo the effect. Done well, it is can be really improve a dull photo. You can actually see some of that effect in many professional photos and HDTV screen images.
 
Unfortunately, unsharp mask tends to increase the "graininess" of an image, i.e. "noise" in areas that are supposed to have a smooth texture. It works best for enhancing soft edges and making them look sharper.
 
Andre knew of a good, free online photo editing program that had an excellent noise reduction filter. "Graininess" often means too much noise. Maybe he'll see this thread.
 
jtbell said:
Unfortunately, unsharp mask tends to increase the "graininess" of an image, i.e. "noise" in areas that are supposed to have a smooth texture. It works best for enhancing soft edges and making them look sharper.
Good point. The OP mentions both low resolution and wanting to "sharpen" them. I would recommend that he find a friend with Photoshop and test a photo to see what will work. Then he can look for those capabilities in free software. I have seen photos where fractal techniques were used to add artificial detail that looked very realistic.
 

Attachments

  • de-jpeg, before-after.jpg
    de-jpeg, before-after.jpg
    34.1 KB · Views: 2,119
Alternatively ''wavelet denoise'' plugin on GIMP , [GIMP and plugins are free] ...
 

Attachments

  • ''wavelet denoise'' plugin on GIMP.jpg
    ''wavelet denoise'' plugin on GIMP.jpg
    27.9 KB · Views: 1,918
Just remember - the image may look better, but as Greg said in the very first answer - it won't have more information than there was in the compressed image. Changes will be only cosmetic, to appease the eye.

Actually in a way they will make the image worse - while it will look better, it will contain other artifacts, not as visible, but still degrading the information stored. You may try to take a picture of a newspaper with different sizes of text (ads, titles, plain text), and then to see which parts of the text can be still read after saving the image with different levels of compression and after applying different filters. My bet is that each additional operation will make the newspaper more difficult to read.
 
  • #10
Borek said:
... My bet is that each additional operation will make the newspaper more difficult to read.

If the brief is to sharpen blurred-text see "SmartDeblur" ...

Before
screen12.jpg


After
resulting.jpg

Now readable, but with plenty of artifacts.
 
Last edited:
  • #11
Interesting and impressive, but that's not exactly the same thing. If the blur is added in a "systematic" way the information is still in the picture - and it is possible to attempt to recover it. When the image was compressed with the lossy compression situation is different.
 
  • #12
Borek said:
If the blur is added in a "systematic" way the information is still in the picture…

It’s me who should say about this. Not exactly as Borek claims because of rounding errors, but blurring an image in computer’s memory, especially if lossy compression is not applied, inflicts much less data loss than blurring due to physical effects in realistic conditions (such as out of the focus photograph).
 
  • #13
Borek said:
Interesting and impressive, but that's not exactly the same thing. If the blur is added in a "systematic" way the information is still in the picture - and it is possible to attempt to recover it. When the image was compressed with the lossy compression situation is different.
A lot of what makes a photograph look sharp is not the amount of information. I think that HDTV looks good even on a very large screen at a size to which I would hesitate to blow up a photograph. If you figure the photograph pixel count that HDTV is equivalent to, it is less than most cameras. Of course an extremely small pixel count will be hard to fix.
 
  • #14
FactChecker said:
A lot of what makes a photograph look sharp is not the amount of information.

Sure - but first, OP never stated what he wants (just an appeasing effect, or getting more information from the picture), second, my bet is OP is not aware of the distinction (so he can be wasting his time trying to do something that is impossible).
 

Similar threads

Replies
1
Views
2K
Replies
16
Views
7K
Replies
4
Views
1K
Replies
1
Views
2K
Replies
6
Views
3K
Replies
7
Views
2K
Replies
7
Views
1K
Replies
7
Views
4K
Back
Top