Working with Core Image


Core Image, in simple terms, puts the power of Photoshop and particularly Photoshop filters into the hands of the average developer. By building a generic suite for working with images, Apple has hugely simplified the way in which we can work with image data. The base of Core Image is a flexible image loading and storage mechanism that allows developers to write software that works with multiple bit depth images. This eliminates the need to worry about working with an 1-bit pixel, 8-bit or even 32-bit per pixel image; Core Image does the hard work for you.

Core Image then builds on this by providing a plug-in architecture called Image Units. Image Units can crop, rotate and otherwise filter the content of the images you are processing. You can even create new filters based on collections of other filters (like running a number of Photoshop filters simultaneously, or sequentially) and you can do all of this at a programming level, making developing software that works with images significantly easier. If the supplied filters aren’t good enough for you, you can also develop your own.

This opens up a whole range of possibilities beyond the usual approach of opening and image and working on it. As a developer you could, quite easily, develop a new droplet to automatically convert or translate an image into another format. A little bit more work (using Folder Actions) and you could make it do this automatically when you drop something on a folder.

But what does this mean for your average joe? Well, the simpler sequences where we’ve needed Photoshop (or GraphicConverter, or similar) may be a thing of the past. I doubt it will wipe out Photoshop; it’s a still a great compositing and image generation tool, as well as one use for manipulating and filtering image content. Photoshop also has a huge range of supplied and third-party plug-ins that will be difficult to beat.

Core Image is part of Tiger, which we all hope will be out soon (I’m betting on a WWDC launch). If you want more info on Core Image and what it can do, check out this great article on Apple’s Developer site.


Jacques Lema

Droplets are a very nice concept. But their goal is to be able to apply _FAST_ a filter or a series of them onto an image. With the years photoshop has become such a monster that it takes ages to load. It’s nowhere near the typical OSX native application.

Imagine i Just want to create a droplet to say resize and crop and image. On my machine 90% of the time will be used by opening and closing photoshop. That’s a big problem. Most native apps, that is apps written with Xcode and using cocoa at 100% launch very fast, even on my slow HD.

I am currently writing an image manipulation program: ( ) and well, it doesn’t have all the features photoshop has (yet :-) , but one thing I know is that if at some point I support droplets (and recorded actions which is the same) the time to load the app will be about 1 second. That makes the difference between usable and enjoyable.

I wonder if we could create a standard (XML based) for applying a suite of filters to an image.


You’re right, you can. But I didn’t say that this is only possible with Core Image.

But the point is that Core Image lets you do a whole lot more than is currently available with Image Events, plus you can easily roll your own modifications.

Kevin Ballard

Actually, you can already use Folder Actions to change the format of images. There’s an Image Events application (or whatever it’s called) that lets you modify the properties of images, and one of them is the format property. I haven’t done this with AppleScript but there’s an equivalent command-line application that lets you change the properties of images and I have changed formats with that.

However, you could write a simple application to apply a series of filters to an image, then use Folder Actions to run that on all images dropped in a folder.

Comments are closed.