Identifying specific characteristics in medical imagery is a type of image processing problem. ANN and SVM can be used to solve image processing problems, typically through MLP type network or SVM. There are good demos in NeuroSolutions software or you can use Matlab's ANN toolbox. An example is given in the following paper. Artificial neural networks in mammography: application to decision making in the diagnosis of breast cancer (1993) Y Wu, ML Giger, K Doi, CJ Vyborny, RA Schmidt, CE. You can locate this paper on Google Scholar!
Each classification problem has its own characteristics, but SVM is usually an interesting option for implementing a robust classifier. I am not extremely experienced in medical images classification, but I have implemented several classification methods for different situations. In fact, I attach you a recent paper where we design a robust vehicle logo classifier by using SVM combined with HOG features. Probably, you can implement a similar approach for your specific problem.
Good luck with your research!
Conference Paper Vehicle logo recognition in traffic images using HOG features and SVM
In image processing and pattern recognition of medical images is a complex task. First you must answer which features you will use. Here many papers and application in med. imaging are available. Please refer to IEEE or ACM libraries. Then you have to decide for a classifier concept. Without a proper definition of good features classifier definition is difficult. I would not only rely on SVM and ANN. In med. imaging Bag-of-classifier-approaches usually help because they can be designed quite robust. Of course an SVM approsch is a good starting point.
In order to familiarize with the basic image classification problem, i would strongly recommend that you try the VLFEAT classification application. You will see the kind of pipelines and representations that are used. You can play around with the code, and maybe even get some initial results for your problem, without too much pain :)