STEPS INVOLVED IN IMAGE PROCESSING
There are many steps that are involved in the processing of an image. Today, we are going to learn all those. Let us begin right now.
There are basically 9 steps that can be performed while processing an image. It is not important to follow all the steps to get a processed image. To know all that is happening in the rear, let us go through all of them one by one.
The steps involved are in order as stated below:
1)Image Acquisition:
Image Acquisition is the first step and the obvious step. In this step, the image is taken in from a storage drive, like a hard disk or any other, into the process. Without the image, there can be no processing done.
Even in the image acquisition stage, there are two sub-processes that take place. They are
1)Image Capturing
2) Image digitization
Image Capturing is the method where the image is obtained from a sensor of cameras or from specific hardware known as frame grabber which can not only grab images but also single video screens or frames to process them and make them high quality. Which when processed, can be installed back in the video. Hence, making the video high quality.
Image Digitization is the process where the captured image’s analog values are converted into a digital value which are of 1’s and 0’s. The media, sound, noise and everything the image holds is converted into digital data so that they can process them easily.
The image is here sampled (breaking down into equal intervals for simple procession) and quantized (compression technique which gathers many values into one quantum value thus helping in compression of the image without losing its quality.)
Now at this stage, after the image is taken in and ready for the process, there is something that is known as Problem Domain.
The term may seem extraordinary but it is not. In simple words, the problem domain shows what is wrong with the image and what needed to be fixed or modified according to the need.
Remember that not all the steps are used while processing the image. Depending on the problem domain, the steps are carried on.
For instance, say that if the image is needing enhancement, then, only the image enhancement process will be used on that image. The methods like morphological processing, segmentation and others will not be used.
2)Image Enhancement:
This is the second step which can increase the sharpness, brightness, pixel density and other factors that can improve the quality of the image. It basically works on two things which are time and frequency.
Imagine a simple bar graph with time on the x-axis(horizontal) and frequency on the y-axis (vertical). Based on these two key elements, the enhancement is done and adjusted.
3)Image Restoration:
In this third step, the image is restored to give it a clean and sharp look. Any distortion like noise, blurry image, misfocus can be fixed using this method.
This step helps to improve the viewing quality of crumbled or degraded images to give it a clean and sharp look once again.
4)Color Image Processing:
This step is only used on the color image as stated by the name. This is used to improve the color vibrancy and viewing pleasure of the image. This process works on the primary colors which are red, green and blue.
During this process, the image lacking any color filter or dullness because of inappropriate color shading or lighting can gain back its purity and look. This is applied on a pixel level, meaning, each pixel is checked and color fixing is done to match the pixel beside, to gain an overall clean view.
5)Wavelets and Multi-resolution Processing:
This is a slightly complex method. In this process, the image is broken down into different degrees and resolutions and then processed, so that they can compress the image without hurting its quality, for the purpose of easy sharing.
The image is broken at different degrees or viewing resolutions and then stored with the intention of smaller size, high quality and faster transfer rate throughout the network.
6)Compression:
This is a simple step where the size of the image is reduced to the minimum so that it can be stored with ease and be shared without any difficulty. This compression is of two types. They are:
1)Lossy: During this compression, the quality of the image may be sacrificed and it may not live up to the expectations of the processed image.
2)Lossless: During this compression, the quality of the image is not at all affected because of the fifth step which is multi-resolution processing. The quality remains the same even after the size is decreased. This is the most preferable method, because who wants to see a low-quality image anyway.
7)Morphological Processing:
This step helps in changing the physical structure of the things present in the image. For instance, there is a person present in the image and the structure of the nose or shape of the face is needs to be changed.
All those changes are done in this process.
8)Segmentation:
It is the process of breaking the image into different small segments or parts mainly on pixel level for simplification or ease of processing.
9)Object recognition:
In this process, the objects in the image, say a person or any other object is recognized based on already present data and the description of that is recorded for future purposes.
For instance, say that there is a tree in the image. The already available data of the tree which says that it will have leaves and branches will be matched and the data will be stored in the database.
So, there you have it! all the steps involved in the image processing. Remember that not all the methods will be used to process the image, only those required based on the problem domain.
Come back for more. Until next time:)
