Welcome to this video. Firstly, let me introduce myself. My name is Consuelo Gonzalo. I am an Associate Professor in Universidad Politecnica of Madrid and my main research area are Image Processing, in particular, with satellite images and medical images. As you can figure out from the title of this video, the main objective of it is to identify the challenges of Big Data in medical images. For that, in the first part of the video, I am going to show you why a medical image analyzes or interpretation can be considered a Big Data. And the question is that most of the V's involved in Big Data problem are present in medical image analysis. After that, I am talking about a semantic interpretation of medical image mean. Also, we are see our first presentation of the image mining overall process and at the end of the video, it will be easy for you to understand what are the challenges. Perhaps, the most important V in the case of medical imagery is the Value. All of you know that this kind of images can be considered it like a window for the human body, since it is possible to obtain crucial information from them. This information allowed to characterize different organs and different diseases. And this information obtained from the images is very helpful for the medical, for the physician to the diagnosis and treatment, different patients and different diseases. For this diagnosis, there are a large number of modalities of medical images. In this slide, you can see a taxonomy just in the case of the radiation used to register these images. But another different criteria can be used to do another different taxonomies. The most important aspect to mention here is that the analysis, the study of different part of the body on different diseases, require different medical image modalities. And in some cases, different modalities should be integrated in order to characterize in adequate way a particular disease. Regarding velocity, in this slide, you can see the figures of ideal summary for a medium size public Spanish hospital. When we obtained information regarding different kind of proofs and test doing in this hospital. From this figure, it is easy to extract the number of medical image, the main speciality of course is that every day are obtaining in this hospital. But the most important thing is not just the number of the medical image registered per day in this hospital. The most important aspect is that all these medical image shall be informed as soon as possible. This is the way that they are really helpful for the physician. The last week V that we are mentioning regarding medical imagery are volume. We have see in the previous slide that this volume per day in a medium sized hospital is very large. And this allow us to get a coarse idea of the number of medical image that are registered every day around the world. But the volume in the case of medical images is not only depend on its number, it depend also of their complexity. They high dimensionality. Most useful medical images are 3D, but we also have 4D and 5D images. And on the other hand, each of these image, half are huge numbers of voxels. Then, if we interpret it, the volume of medical image in terms of voxels, the complexity of the analysis of this kind of data is equivalent to the complexity of genomic data analysis. We can talk about more of these in the case of medical images but I think it is enough with these four Vs that we have discussed before to be sure that the medical image analysis is a Big Data problem. Then in this situation, what we need is a techniques that allow us to extract information from the images from this huge volume of data that we mentioned it before, that we know that are register it or generate it in a really fast way and that they present a high variety and velocity. These techniques today are under the umbrella of Big Data. However, there are another areas of the science and engineering that also addressed this kind of problem. And then, in this case, we can talk about image analytic, image mining, image understanding. But in all the cases independently on the name that we use to define them, the final objective is provide techniques to the physician that allow them move toward a evidence-based medicine. This is not a trivial problem. Why? Because in the images, what we observe when we have our image to be analyzed, are the values, shapes that store sometimes relation between object context but when the physician see its diseases, organs, etc. The difference between these two kind of ideas/concepts is what is named Semantic Gap. Then other work to say what we need to extract information, knowledge from the images is we need techniques to treat this Semantic Gap. There are different approaches in the literature. One of them is Image Mining. This is the approach that will be developed in the next video in the model of medical image analysis. In this future videos, we describe the overall process of image mining. You have here an scheme of it. And I think from this previous idea, it is easy for you now to understand that the main challenges of Big Data in medical image today is try to develop techniques that finally provide tools to the physician for a detect in automatic way, human organs, pathologies that they help to the physician to detect or to quantify different diseases, like for example, the stage of cancer. Also, a physician need tools to do temporal evolution studies for his patient. And at the end, what we need is that computer help a physician to understand, to interpret medical image in a faster way than now they can do and also a more precise way. You can find a lot of information regarding this issue in Internet or in a scientific journal, a magazine, etc. However, I will recommend read this to a preference that I introduced in these slides. Thanks for your attention. I will go and see in next videos.