CHAPTER 3
Overview of Image Registration Techniques
The field of Image Registration is an immense and ever expanding field. By the early stages of 1993, There existed over 120 papers written on registration problem, as cited in survey article written by van den Elsen et al.[199]. Since then the number of papers published have grown exponentially.
This chapter will discuss elements of registration techniques according to a classification that was originally proposed by van den Elsen et al.[199] and later extended by Maintz et al .[120]. The set of criteria described is explained in fig 3.1.
This classification includes algorithm’s dimensionality, nature of registration algorithm, nature and domain of transformation, user interaction, optimization procedure, modalities involved and type of subjects used in algorithm.
.
3.1 Dimensionality- 2D, 3D, 4 D
One of the most obvious classifications which came out from the set of image registration technique is about Dimensions that means how many dimensions are used in the registration process. The range of this dimension can be from a simple 2D to a complex time series registration of 3D data, i.e. 4D process [49].On the basis of dimension registration algorithm can be decided into two parts-those that deal with time series registration and those that do not,i.e. only deal with spatial dimension.
3.1.1 Registration involving Spatial Dimension
The algorithm that only deal with this field of spatial dimensions can be further classified into 2 D and 3D. Registration algorithm can be applied to both 2D and 3D data sets very easily. The only difference in 3D however, is that the size of data set is greatly increased and the number of transforma...
... middle of paper ...
... tomographic images are matched with an anatomical atlas or some other model. Such procedure can facilitate automatic segmentation [42].
2.7 Subject
This classification refers to subjects that are involved in registration process. This can be classified as inter-subject, intra-subject or subject to model registration. Intra-subject is mostly used .In this data is from same scenario, like from same patient. This type of registration can be used in almost every diagnostic.
Inter-subject is quite complicated as transformation must overcome the inherent anatomical differences that exists between two different scenario. That’s why mostly inter-subject algorithm is based on curved transformation.
Subject to model is essentially the same as that of modality to model. Thus all the techniques for modality to model is also applicable to subject to model registration also.
Three dimensional motion capture requires more than one camera to create depth in the motion being performed. A good example is eyesight. If someone where unfortunate enough to only have one eye they would be unable to see the 3-D depth in motion. Having two eyes allows for depth in motion when seeing, which is similar to the idea of using two cameras in order to fully capture the depth in motion. One of the few techniques discovered and used in 3-D motion capture is Direct Linear Transform (DLT). Using the idea that, images from the cameras are determined by their placement to discover their distances, equations could be formed and used. Test subjects wear reflective markers to allow the cameras to follow their movement and motion through space. These reflective markers are placed on certain joins and parts of the body the researcher would like to study. The reflective markers ca...
...ge flow and pattern types, are prominent enough to align fingerprints directly. Nilsson [26] detected the core point by complex filters applied to the orientation field in multiple resolution scales, and the translation and rotation parameters are simply computed by comparing the coordinates and orientation of the two core points. Jain [27] predefined four types of kernel curves:first is arch, second is left loop ,third is right loop and fourth is whorl, each with several subclasses respectively. These kernel curves were fitted with the image, and then used for alignment. Yager [28] proposed a two stage optimization alignment combined both global and local features. It first aligned two fingerprints by orientation field, curvature maps and ridge frequency maps, and then optimized by minutiae. The alignment using global features is fast but not robust, because the
...omated detection of lines and points in the images and the use of smart markers in reference video recordings.
In 3D ultrasounds, the computer takes multiple two-dimensional images at various angles and arranges them to form a three-dimensional
T. Sielhorst, T. Blum, and N. Navab, “Synchronizing 3d movements for quantitative comparison and simultaneous visualization of actions,”in
from the same problems of a computer based system, such as illumination, occlusion and pose variations,
Scaling is a part of geometric transformation. Scaling transformation is used to change the size of an object either to shrink or enlarge (Yuwaldi, 2000). Object scaling is normally used in computer graphics. For example, the user can enlarge or shrink the drawn object according to certain specifications. In the medical field, scaling techniques are used by experts during the pre-surgery process. For manual template method, the surgeon must face two different expansion image i.e scale of the X-ray and the implant (template) scale (Siti Fairuz, 2009). The surgeon takes longer time to identify the appropriate implant size due to different resolution in the patient’s X-rays (Fang et al, 2006). This research is conducted with the purpose to show techniques and algorithms that can solve the problem of scaling in the medical field that involves the use of medical images and digital implants.
This approach includes two processes, training and classification (Chelali, Djeradi & Dejradi, 2009). In the training process, a subspace will be established by using the training samples, and then the training faces will be projected onto the same subspace. In the classification process, the input face image will be measured by Euclidean Distance to the subspace, and a decision will be made, either accept or reject.
[Jain, 2004] Jain, A.K.;Ross, A.;Prabhakar, S.;"An introduction to biometric recognition", Volume: 14 Issue: 1 Issue Date: Jan. 2004, on page(s): 4 - 20
“Virtual Humans are artificial agents that include both a visual body with a human – like body and intelligent cognition driving action of the body” (Traum, D., 2007). It can have many roles such as acting as a role player in a training system, acting as a tutor, and even have a role in a game. These virtual humans can be used in many different field of work. Nowadays, people even used the virtual humans as a medical application. The previous one was involved with PTSD and ADHD that use systems with virtual reality. Other than that, the virtual humans also can identify the ethnics and cultures. There are different in their conversational behavior of virtual agents. In this intelligent, they also use many techniques to make the virtual agent look real. They can create virtual humans with the natural gesture and face expression. It also can make an emotionally expressive head and body movement. Through these things, we can start to recognize the functions of the virtual human model.
raw hand data is to combine the previous two methods in a hybrid approach with the
Images of human anatomy have been around for more than 500 years now. From the sketches created by Leonardo da Vinci, to the modern day Computed Tomography (CT) or Magnetic Resonance Imaging (MRI) scan, images have played a great role in medicine. Evolution in medical imaging brought together people from various disciplines such as Biology, Physics, Chemistry and Mathematics, a collaboration which has further contributed to healthcare as a whole. Modern day imaging improves medical workflows by facilitating a non-invasive insight into human body, accurate and timely diagnostics, and persistence of an analysis.
Minutiae-based techniques: In these minutiae points are finding and then mapped to their relative position on finger. There are some difficulties like if image is of low quality it is difficult to find minutiae points correctly also it considers local position of ridges and furrows not global [4].
Virtual reality usually used to describe some collection of technology devices that has abilities to communicate interface based on interactive using the 3D visualization and suitable to collect the databases with the natural skills and senses like a single real but in different inputs are provided. For example, 3D visualization from the computer have been to display all the data gloves using one or more position while can help to head mounted will displays and then an instrumented clothing like suits and gloves is less than 20% of virtual reality application in medical are used in the immersive equipment, however an attention and behavior of sciences know where immersive devices can detect more than 50% application. It is can explore and explain how far the advantages and what is the great advances in human computer interface allow the user to interact with technologies, (Rubino et al, 2003). Besides that, virtual reality also has been to represent of virtual object with human senses while identical to their natural counterpart. These medical technologies were become more information based and it will be help to represent a patient with higher fidelity to a point and that can detect any an images such maybe a surrogate for the patient or called any of the medical avatar. For example, some of the more great effectiveness technologies in virtual reality system looks like a real part of the body or that avatar can interact with external devices like surgical instruments, (Jones,2003). In addition, virtual reality has provides a new version of human computer interaction paradigm especially in clinical specialists the ultimate goal. That will be to simply external observe of images on a computer- generated three dimensional virtual wo...
Machine learning systems can be categorized according to many different criteria. We will discuss three criteria: Classification on the basis of the underlying learning strategies used, Classification on the basis of the representation of knowledge or skill acquired by the learner and Classification in terms of the application domain of the performance system for which knowledge is acquired.