image processing image representation1.pptx

Advance computer vision unit 1

1. Image representation in computer vision refers to the process of converting an image into a numerical or symbolic form that can be easily understood and processed by a computer. Images are typically represented as a collection of pixels, where each pixel corresponds to a specific color or intensity value

2. Mathematical Model used to describe image and signal A signal is a function depending on some variable with physical meaning(temp,pressure distribution,distance from the observer etc) It can be one dimensinal(depending on time) two dimensinal(depending on two cordinates o the plane) three dimensinal(describing volumetric object in space etc) higher dimensinal Scalar function might be sufficient to describe monochromatic image vector function for image processing to represnt (eg color images consisting of three component colors)

3. Function 1)continuous : has continuous domain and range 2) Discrete or digital:domain is discrete and range also discrete Image on human and tv retina can be model as function ot two variable f(x,y) where x,y are coordinates in plane or perhaps three variable (x,y,t) where t is time

4. image is aquired in many ways normally color is norm althouhg we present algorithm for monochromatic images 1)cameras operate in infra red part (for night surveillance) 2)electromagnetic spectrum 3)teraherts imaging 4)also outside the EM spectrum (light) is also common in:A)medical domain,data sets are generated through magnetic resonance (MR), B)computed tomography (CT) C)ultarsound etc.

5. all these these opeation generate huge data ,it requires analysis and understanding and with increasing frequecy these array are of 3 of or more dimesions.

6. continuous image function gray -scale image function values correspond to brightness at image point. the function value can express other physical quantities as well (t emp,pressure distribution,distance from the observer etc ) Brightness intergrate different optical qauntities . Image on human eye and TV camera sensor is in 2D can call this as intensity images(bearing information of brightness) 2D image is result of projection of 3D scene. 2D intensity image is result of perspective projection of 3D scene (though pinhole camera) A non linear perspective projection often approximated by a linear parallel(orthographic )projection

7. image Digitization image to be processed by the computer must be presented using disrete data structure eg a matrix. An image is captured by a sensor is expressed as a continuous function of f(x,y) is sampled into a matrix with M rows and N columns . Image quantization assigns to each continuous sample an integer value -continuous image function f(x,y) is split into k interal s. ( Nyquist criterion requires that the sampling frequency be at least twice the highest frequency contained in the signal, or information about the signal will be lost.) the finer the sampling (i.e the larger M and N) and quantization (the larger k),the better approximation of the continuous image function f(x,y) achieved.

8. image sampling poses two:1)sampling period should be determined (this is the distance between two neighboring sampling points in the image) 2)geometric arrangment of sampling points (apling grade should be set )

9. sampling

10. sampling

11. low level processing : has very little knowledge about the content of images low level methods often include image compression ,pre processing methods for noise filtering ,edge extraction and image sharpening It takes input as image captured by the a TV camera is 2D in nature.

12. In digital image processing and computer vision, image segmentation is the process of partitioning a digital image into multiple image segments, also known as image regions or image objects (sets of pixels).

13. High level processing: is based on the knowledge ,goals and plans of how to achieve these goals adn Artificial intelligence methods are widely apllicable. high level computer vision tries to imitate human cognition and the ability to make decisions according to the information contained in the image. high level vision begins with some form of formal model of the world , and then the reality perceived in the form of digitized images is compared to the model .

14. A matched is attempted ,and when differences emerges,partial matches (or sub -goals ) are sought that overcome the mismatches;the computer switches to low level image processing to find information needed to updates the model. this process is repeated iteratively ,and ‘understand’ an image thereby becomes a cooperation between top-down and bottom - up processes. a feedback loop is introduced in which high -level partial results creates tasks for low-level image processing ,and iterative image undrstanding process should eventually converge to the global goal. computer vision expected to solve very complex tasks ,the goal being to obtain similar results to those provided by the biological systems.

15. Continued with low ,high level processing