Editing photo images by computer
The various steps
Editing photo images by computer entails:
- Transforming a negative or slide into a digital image, unless of course one uses a camera that stores images in digital form.
- Editing the digital image.
- If you wish, to apply a screen.
Digitizing an image
Prior to computer processing an image, it must be scanned. You can use a black-and-white or colour scanner to copy pictures and book illustrations. There are special scanners for slides and colour or B/W negatives.
A scanner is an instrument with which one illuminates an image with very narrow beams of light. A photoelectric cell transforms the reflected light into a current, which in turn is transformed into a signal by some piece of electronics or a computer. The entire image is illuminated and scanned by moving the light beams in very small steps. The size of the steps with which the beam is moved on a line, and the distance between successive lines, determines by how many points (pixels) per cm or per inch the image is being represented. The usual resolution of popular scanners is 300x300 or 600x600 dpi (dots per inch); this resolution may be artificially improved by interpolating between adjacent points. Further technical details are beyond the scope of this communication.
In general it is best to store images with the maximum information:
- Perform scanning at high resolution (dots per inch). You can later reduce the information content to a minimum; conversely, if you start with too low a resolution this cannot be redressed. Many software packages let you specify the resolution at which the image should be stored: select a relatively high resolution.
- Colour photos can be stored in black and white; an advantage is that this greatly reduces file size. It is often preferable to first perform colour correction and then transform the image to B/W. You may be inclined to store scanned images as GIF or JPG files. The obvious advantage here is that the file size is significantly reduced by employing algorithms that discard information that is redundant to the human eye. However, eliminating redundant information is best done during image manipulation, and not prior to it.
- Store the image at the size at which you will ultimately use it, or larger. If you have to later increase the image size new points need to be created by interpolation; conversely, when you decrease the dimensions points need to be discarded. The algorithms used in such processes lead to jagged lines, in particular if the initial resolution was poor so that the image was represented by only a limited number of pixels.
Manipulating a digital image
The objective of manipulating the image is to obtain a picture with sufficient details or lines to reproduce the original image, but not so detailed that it cannot be reproduced by casting. Ultimately the image must be reduced to pure black and white. It follows that we should remove most of the information available in the original image, such as for example colour. You should therefore have some preconceived idea of what you want to achieve; creativity is required to reduce the image to its essence. You should also develop experience in manipulating digital images by computer. Here are some general guidelines:
- Particularly when you want to cast thin objects, such as with the Delft casting method®, the objects need to be small, say 4x4 cm; casting larger thin objects without access to professional equipment is much more difficult.
- Start by cropping the original image to its final dimensions. If you wish you can also correct the colour and tonal values; then store this image. If you subsequently manipulate the image but are dissatisfied with the results, you can always start from this original again. It is a matter of personal preference whether you store in full colour or as a B/W image.
- If you wish to transform the bitmap into a scalable (vector) format by redrawing it, so that the image is now made up of lines and areas, start from the original image. Vector images can be enlarged or reduced in size at any time without any loss of detail; this is because the (bounding) lines are transformed into mathematical equations, so that they can be scaled to any size. If, however, you wish to manipulate the bitmap, then scale it to the size that you will ultimately use at an early stage, and work on this image.
- The ultimate bitmap image will be more photorealistic the more detail (pixels) is preserved. However, part of small details is lost in transferring the image to the mould (photo polymer), subsequently upon casting, and finally in finishing the cast object. Therefore, by all means eliminate as many details as possible, preserving both the quintessence of the image and making it castable.
- In contrast to rasterisation described above, where black points are commonly separated by a white zone, in a black and white bitmap each bit is transformed to either black or white. Dark areas therefore lose all detail and merge with black ones, so that the result may be dull. Light areas similarly become optically empty. Do take this into account when selecting an image. The bitmaps (rasterised images) in figure 6 cannot be reproduced upon casting, with the exception of the third bitmap. At 40 dpi there is loss of detail due to adjacent black pixels 'merging', and the large number of black dots makes the image grubby.
- Arising from this try to achieve a pronounced black and white contrast for strong, graphical images. Avoid small black dots in large white areas (or white dots in black areas): they will look like a contamination. Critically inspect the image after printing on paper, and touch the image up prior to using it for the casting process.
The following example illustrates some of the manipulations required. In this example a limited number of manipulations sufficed, and the same results can also be obtained by different procedures; often more manipulations are required, using filtering and other facilities more extensively.
Example of editing a digital image
A colour scanner was used to scan in a colour picture. From then on different strategies can be adopted, and there are probably more efficient ways to achieve the end result. We used Adobe® Photoshop® to edit the image as follows:
• Colour correction (not shown)
Convert to greyscale, use magic wand, lasso tool and pen tool to isolate the child against a white background. By the way, with Photoshop version 6 isolating a part of a picture from unwanted background is much easier using the Extract feature. |
 |
Use the lasso tool to isolate a detail and modify the contrast. |
 |
Enhance the contrast of another area after isolating it with the lasso tool. |
 |
'Filter', |
 |
'Image', |
 |