Skip to 0 minutes and 5 secondsThis course is about image analysis methods for biologists. Image analysis is a branch of computer science in which we're concerned with taking digital images of the world and extracting, from those images, some kind of quantitative data that describes the objects and the things we see. With the recent increase in demand for plant phenotyping, especially automated plant phenotyping using robotics, there's an increased interest and a need to measure things automatically in these images.
Skip to 0 minutes and 34 secondsAnd so this course is designed to give you an introduction to things you need to think about when you're capturing images and how to start to go about analysing them and looking at some of the techniques that you can use to start to get at the interesting data in your images. Why are we doing this for biologists? Well, images are everywhere in biology these days. Biologists typically use colour cameras, microscopes to look at populations of cells dividing and growing, specialist devices like confocal laser microscopes to make three dimensional images of the structure of the samples, and more recently, things like microcomputer tomography, x-ray machines, and magnetic resonance imaging to look at the 3D structure of larger objects.
Skip to 1 minute and 17 secondsWe'll look at some of the things to think about during image acquisition, how do take the best quality images, what to do with your images if they are still affected by things such as image noise, reducing the overall quality of the image. And then once you've got those images, you want to do things like identifying which pixels belong to plants. It is possible to take these images and manually mark them up to have a user look at them and point at the points of interest and make measurements by hand. The difficulty with that is that the users tend to get very tired, very quickly.
Skip to 1 minute and 51 secondsThere tends to be little variation between the measurements produced by different people, and overall, the data that you produce is quite subjective. It also takes a very long time to do. What image analysis methods can do is provide automated methods, in the form of software tools, that can take an input image and automatically objectively produce accurate quantitative data with a minimum of human intervention. And so this course should give you a good overview of some common techniques that you will use, perhaps where the future is heading as well, and we'll try to give you some pointers to some more advanced topics, as well as covering the basics.
Skip to 2 minutes and 30 secondsWe don't assume that the people watching these videos are computer scientists or have any prior knowledge of image analysis and computer vision. We're not aiming for a very detailed understanding of the techniques, just enough to allow you to exploit the methods that are already there.