computer vision

computer vision

Computer vision can be simply defined as the way by which computers are made to get a high-level understanding from images and videos. It can do tasks which are similar to what human visual system can do. Computer vision tasks include the production of numerical information in the form of decisions by analyzing the digital images and extracting the data from the real world. This image can be understood by using models which are constructed with the aid of physics, geometry, statistics, and learning theory.

There are two disciplines with which computer vision deals; scientific discipline and technological discipline. The extraction of the information from images which is done by the artificial systems, this is what computer vision does as a scientific discipline. The application of theories and models to construct a computer vision system is considered as a technological discipline.

How computer vision works?

Our brains recognize the images as our eyes see them. So, when we look at any picture, we can describe what we see. We can recognize colors, for example, we can know that this thing is a tree and its color is green. That's for the human being, but what about computer vision?

Computer vision works with numbers. There are many functions which are similar in many computer vision systems:

ü Image acquisition. Several image sensors besides several types of light sensitive cameras are used to produce the digital image. The resulting image data will be an ordinary 2D image, a 3D volume, or an image sequence Depending on the type of sensor.

ü Pre-processing. It's so important to assure that the data satisfies the assumptions implied by the applied method so, before the application of any computer vision method to image data, it's necessary to process the data. For example:

-        Re-sampling.

-        Noise reduction.

-        Contrast enhancement.

-        Scale space representation.

ü Feature extraction. The extraction of various image features from the image data, these features such as:

-        Lines, edges, and ridges.

-        Points, corners, and blobs.

-        Texture, shape, and motion.

ü Detection/segmentation. This step concerns with the further processing which needed for some points or regions of the image. For example:

-        Selection of specific points.

-        Segmentation of many image regions.

ü High-level processing. At this step the input is considered as a small set of data and the remaining processing deals with:

-        Verify the satisfaction of data for the assumptions of application.

-        Estimate the specific parameters of the application such as, pose and size.

-        Image recognition.

-        Image registration.

ü Decision making. In this step, the final decision of application is made, for example:

-        Pass / fail on automatic inspection applications.

-       Match / no-match in recognition applications.

Computer Vision Applications

§  It's used widely for automatic inspection such as in manufacturing applications.

§  It's used as an assistant for humans in identification tasks.

§  Controlling processes.

§  Medical image analysis and processing.

§  It's used widely in navigation for the design of robots.

§  It's widely used in the medical field to diagnose patients.