SA :)
This is a summary for a paper that proposes an automatic computer-aided diagnosis (CAD) system
for early detection of lung cancer by analyzing chest 3D computed tomography (CT) images. 
This is a very useful paper, it discusses the steps followed by the system and states the algorithms used briefly.


For the summary : Click here
For the original : Go here and then download the full issue, the paper's name is:
"Computer Aided Diagnosis System for Early Detection of Lung Cancer Using Chest Computer Tomography Images" 


It's a good site by the way with many papers available for free :D

السلام عليكم و رحمة الله و بركاته

Before i begin talking about SNAKES, first i have to answer some questions.

What is image segmentation?
- It's subdividing or partitioning an image into it's constituent regions.

Why?
- To extract features of interest in images by looking for mathematical entities which describe the shapes of objects appearing in images.

So, why do we need image segmentation in our project?
- we want to extract the lung area from a CT (Computed Tomography) image which is our interest, for computing and detecting the abnormal shadow areas that are suspicious to be tumors.

What we have to do?
-we have to find a method that looks for any shape in the image that is smooth and forms a closed contour.
-From here we begin our research in "SNAKES".

Introduction:

• The active contour model algorithm, first introduced by Kass et al., deforms a contour to lock onto features of interest within an image.
• Usually the features are lines, edges, and/or object boundaries.
• Kass et al. named their algorithm, “Snakes” because the deformable contours resemble snakes as they move

Framework for Snakes

• The active contour system provides the ability to interactively guide the contour detection by placing initial points
• A higher level process or a user initializes any curve close to the object boundary.
• The snake then starts deforming and moving towards the desired object boundary.
• In the end it completely “shrink-wraps” around the object.

CBIR .. what does this stand for ?!
CBIR stands for Content Based Image Retrieval.
and in brief CBIR is a technique used in image retrieval from large image databases, instead of relying on on metadata such as captions or keywords in retrieving images which is expensive while saving images and inaccurate in retrieval.
"Content" may be ambiguous, to be clear content of an image may be colors in it, textures, shapes or any other information that can be extracted from an image.


The following figure shows a general structure of CBIR systems 


CBIR systems always consists of 2 phases the offline phase and the offline phase



where “offline” indexing phase is displayed in the bottom part. Usually the visual contents of the images in the database are extracted and described by multidimensional feature vectors. A feature database is formed using the feature vectors of the images in the database. The “online” content-based retrieval is displayed in the upper part. Users provide the retrieval system with query example images, which are used to retrieve images.
The system then extracts the feature vectors from the example images. The similarities/distances between the feature vectors of the query example and those of the images in the database are calculated. The online and offline phase interact with a collection of multimedia items (images, videos, etc.) from a multimedia database. The query provided by the user can be an example image, region, sketch, humming, or text .


Medicine is one of the most potential application area.

With medical imaging techniques such as X-Ray, computer tomography, magnetic resonance imaging, and ultrasound, the amount of digital images that are produced in hospitals is increasing incredibly fast. Thus the need for systems that can provide efficient retrieval of images of particular interest is becoming very high.

Since for very large database personally describing and annotating every image with text indices is time-consuming and impractical (TBIR - Text Based Image Retrieval or also called Metadata approach), CBIR seems to be a good approach.
Researchers are still working in this field, it's a new born field.
and thus we'll participate in researches and applications in this field isA :)

Resources:
1- Ivica Dimitrovski, Dejan Gorgevik, Suzana Loskovska, WEB-BASED MEDICAL IMAGE RETRIEVAL SYSTEM


2- http://en.wikipedia.org/wiki/Content-based_image_retrieval

السلام عليكم ورحمة الله وبركاته