View allAll Photos Tagged imageprocessing

WFPC2 450W, 606W, 814W

 

Robert Gendler

email: robgendler@att.net

Website: www.robgendlerastropics.com

 

photoshop; photo retouching; retouch; photoshop editing; photography; background removal; adobe photoshop; photoshop work; amazon listing; color correction; photoshop editing; photo manipulation; amazon product photo; photoediting; background remove; image editing; amazon; remove background; retouching; image resizing; photo edit; amazon photo editing; background removal; image resize; ebay; cut out; online shop; photoshop image editing; resizing images; fashion; amazon store; resizing; jewelry edit; photo manipulation; crop; remove background; resize image; fashion; photoediting; resizing; transparent; photo manipulation; amazon; photoshop; white background; retouch photo; changing background; product retouch; ecommerce product; photoshopping; photoshop work; clipping; product editing; photoshop work; photo editing; photo manipulation; amazon infographic; cut out; amazon editing; cut out images; photoshop retouch; cut out background; photo fixing; resizing; e commerce; photo resize; resizing images; resize; image processing; nick add; image masking

differen ways of image processing.

Taken from the castle. HTC One X Panorama function.

Extremely abstract and highly saturated version of a glitched and remixed painting by Tiziano Vecellio, the result of various processes, including an FFT that cut off high-frequency detail.

 

Animated GIF version here.

Created with GlitchSort. Remix and mashup of commons.wikimedia.org/wiki/File:SundaySchool.jpg, by Tom Mandel, licensed under the terms of the cc-by-sa-2.0.

Tool: Ribbet

Effect: Circle Splash

Creator: Jerry Lin

Settings: Normal, Fade 0%

Museum of Flight. East Fortune.

TIred of prounouced sunbursts, I openened for this sunset to f11 from my usual minimum f16.

Image processing stages - left to right, top to bottom:

 

1. raw image from registax6 immediately after stacking

 

2. registax wavelets - settings: gaussian wavelets, increment 1, wave 2 slider slightly negative, wave 3 slider range 4-10

 

3. registax6 histogram and colour balance: use the histogram slider to adjust the midpoint, and also stretch the dynamic range (log scale view to avoid clipping).

 

Use colour balance panel to turn down the red and blue channels for correct-ish eyeball colour.

 

4. restart registax6 and import the saved .png image. Gentle use of wavelets, same settings again. Wave 3 slider range 2-4

 

5. registax6 histogram and colour balance: same procedure as before

 

6. import image into Photoshop elements. Gaussian blur at minimum settings (0.4/0.5) to remove noise and obvious pixelation/TV scan lines.

 

3 rounds of High Pass Masking (overlay setting - hard light setting is too fierce, soft light too modest) and if necessary a bit of Gaussian blur if speckles start to intrude.

 

No idea what HPM does - works like a much gentler version of unsharp masking, and only 1 parameter to adjust !

 

7. More Photoshop Elements: adjust levels RGB midpoint for brightness, Red and Blue channels for colour balance again.

 

Visualising SIFT descriptors (on frames from Wall-E) more description of the method (using openCV and Python) at my blog: www.cutsquash.com/2014/11/visualising-sift-descriptors/

After Quentin Massijs

Detail of Little Boy 006. Glitch, remix of commons.wikimedia.org/wiki/File:Atombombe_Little_Boy_2.jpg, an image of the Little Boy atom bomb. Here ithe image has been eroded until it disappears.

 

The color effects of the image glitching/erosion method I'm using are somewhat more apparent here than in others. the colors seem to sum to gray when high resolution images are reduced in size.

Experimenting with colour versions of the classic image processing technique of the Hough transform.

See more details and examples here:

www.cutsquash.com/2014/02/colour-hough-transform/

 

Code available here:

www.openprocessing.org/sketch/134254

What is that feeling, when everything is too big and learning to play is a risky adventure?

Glitch and remix of a painting by Tiziano Vecellio (Titian). The digitized version had highly saturated color, making it an interesting candidate for color and detail degrading with an FFT. The degrading of detail is especially apparent here.

The infamous drone, after a makeover.

Technical note: Noise filtered with an FFT at various frequency bands, with sorting, quantization, channel-swapping and some masking/remixing and editing tricks. A heavily glitched image intervened to create the lighter and darker areas and some of the color, otherwise it had little effect.

To be gazed upon from a distance, until the Mage appears. Then run as fast as you can.

resize image website: imageupscaler.com

Image Upscaler - resize image without losing quality

 

Image Upscaler, an online photo enlarging service, helps you resize images without losing quality. Using artificial intelligence, it increases the resolution of the image saving its natural look. If you don’t want your enlarged image to be as appealing as a second-hand burrito, keep reading.

 

Who needs Image Upscaler

 

In fact, for the last years, photos are rocking the world. That’s why a picture resizer is just a must-have tool for:

 

• Instagram, Facebook, Twitter, Pinterest bloggers;

• websites owners;

• self-development coaches;

• students, managers, and others;

 

Don’t waste your time trying to enlarge a picture being devoted to a usual photo editor! The new app can cut your image keeping it clean from large pixels.

resize image website: imageupscaler.com

Process: I filtered out high frequencies ("details") using a statistical FFT at different scales on each color channel (RGB), then I attempted to recover the image by enhancing the remaining high frequency data (using a statistical FFT with different settings), but pushed the recuperation process beyond the limits of "good" imaging.

Soldiers who hunkered down in the trenches at the Nevada Test Site pose for a picture.

Mostly a product of Fast Fourier Transforms on different channels in HSB and RGB color spaces, starting from an image of Luther Blisset.

1 2 ••• 15 16 18 20 21 ••• 64 65