MEDICAL IMAGE ANALYSIS

MEDICAL IMAGE ANALYSIS - INTRODUCTION

Image processing is any form of signal processing for which the input is an image,
such as a photograph or video frame; the output of image processing may be either an image or, a set of characteristics or parameters related to the image. Most image-processing techniques involve treating the image as a two-dimensional signal and applying standard signal-processing techniques to it.

The acquisition of images (producing the input image in the first place) is referred to as imaging. Imaging has become an essential component in many fields of biomedical research and clinical practice. 

Medical imaging is the technique and process used to create images of the human body (or parts) for medical procedures seeking to reveal, diagnose or examine disease) or medical science (including the study of normal anatomy and physiology). Medical images are the basis of diagnostics, treatment planning, and treatment. They are also important for medical education, research, and epidemiology.

The discipline of medical image processing deals with generation and reconstruction, pre-processing and improvement, analysis and quantification, as well as visualization and management of all kind of medical images. Image processing techniques serve three purposes in medical diagnostics:

1. Reconstruction of images acquired using tomographic techniques.

2. Improvement in the appearance of images for viewing.

3. Preparation of images for quantitative analysis.

The goal of medical image analysis techniques is to perform automatic or semi-automatic (interactive) medical diagnostics. Medical Image Analysis helps to extract information from image data (Morphometry, Visualization, Monitoring, Functionality, Image-guided Planning, and Therapy). It is important to note that abuse of image processing techniques can introduce features (artifacts) that are not part of the inherent image information. Such artifacts can obscure, or be mistaken for, diagnostic features. This can result in misdiagnosis. 

Medical Image Analysis - Functional Blocks

Functional Blocks of Medical Image Analysis

Image Acquisition:

The images of the human body or its parts are to be acquired for medical procedures that reveal, diagnose or examine disease. As previously mentioned, the images can also be used in the study of anatomy and physiology. The few types of medical images are,

1. Computer Tomography (CT)

2. Magnetic Resonance Tomography (MRT/MR/MRI)

3. Positron Emission Tomography (PET)

4. Single Photon Emission Computed Tomography (SPECT)

Types of Medical Images

Image Segmentation:

Image Segmentation is the partition of the image into homogeneous regions with small gradients and connectedness that correspond to anatomical structures. The medical purpose of image segmentation is the identification and classification of organs or tumors and quantify it (area, diameter, or volume measurement, particle count, length or size distribution).

The most common approaches for image segmentation are Region Of Interest (ROI) tracking, clustering, texture analysis, statistical grouping, and spectrum analysis. Several semiautomatic approaches have been devised, but up to now, automatic segmentation is a very difficult problem. 

Image Registration: 

This technique helps to temporally overlay either same modality or different modalities of medical images. This enhances the analysis part due to integrated features of more than one modality like SPECT & CT or MRI & CT fusions. This step is most advantageous for fusing anatomical and functional images. The most common methods for image registration are,

1. Matching with point-based methods: The image planes are mapped according to their co-ordinate relations.

2. Matching with surface-based methods: The closest regions are identified based on their feature similarity measures.

3. Matching with intensity-based methods: The images are grouped based on their grayness or color distribution.

Visualization:

Visualization gets the information from medical image data using medical image analysis. This helps to model multi-planar information using 3D object modeling techniques. Thus, the image data is transformed as 3D models through volumetric deformation. The tools like 3D slicer are more powerful to display the 3D medical images for interoperative human expert visualization. Also, for the surgical purpose, this technique provides depth information of tissues, functional parts of patients and provide more understanding during diagnosis. 

The applications of Visualization are:

1. Coordinate visualization with intraoperative instruments

2. Surgical Planning & Simulation

3. Maximize Tumor Removal

4. Minimize Damage to Critical Structures

Interpretation:

Interpretation

This stage helps to provide meaningful interpretation for understanding the objectives correctly. This can be done either quantitatively or qualitatively. The metrics for validating the outcome differs based on the application and depth of information in interest.