INTRODUCTION
1.1 History of Mosaic Image
A mosaic is exceptionally simple in concept it is created by combining various colored smaller pieces, called tessera, into a larger added significant overall shape. In ancient times, these works were often made from colored ceramic or stone tiles, decorating floors and walls and creating stunning figural compositions of everything from great battles to mythological scenes. Over time, the concept evolved into more complicated forms, such as the portraits created by 16th Century painter Giuseppe Arcimboldo.
In his work he had used ordinary objects such as fruits, vegetables, flowers, fish, and even books in such a way that the whole assortment formed an identifiable likeness to the portrait subject. Soon the idea of using other images as tessera emerged leading the way to works such as the 1976 Salvador Dali portrait of Abraham Lincoln.
He composed this work by combining images with his own unique visual content, into a realizable portrait of the former U.S. President.
Figure 1.1: Mosaic Image of Abraham Lincoln
In the mid-1990s, this idea of using other images as the tessera for a larger mosaic image sparked an idea in mind of MIT graduate student Robert Silvers. An automated computer system was devised by him so that it that would take any image as input, and produce as output an impressively accurate visual representation of that original input built entirely from smaller equally sized photographs.
These computer generated image mosaics, or Photo mosaics as Silvers called them, soon found their way into homes in the form of posters, jigsaw puzzles and various other 2 forms of merchandise. Despite Silvers holding patents on his process in several countries, dozens of software packages soon began to surface that would allow even the most novice users to create their own image mosaics from their digital photo collections.
Figure 1.2: Lighthouse image mosaic created by Robert Silvers’ system
There has been exponential progress in the modern history of mosaic display. In n-way tables, it has progressed from simple plots of observed frequencies, to mosaic a plot which shows that in log linear model; there is the lack-of-fit, which has progressed to interactive systems that is providing visual fitting along with exploration. As a result, there may arise interesting questions in future due to these significant but exponential developments for categorical data in this field, traditionally the poor cousin of quantitative graphics. These include questions like: for categorical and quantitative variables, plotting matrices for mixtures. And comment on marginal vs. conditional views (Friendly, 1999). We may also look forward to the increasing development of interactive and dynamic methods for exploration, model specification, fitting, and diagnosis with categorical data.
The applications of the space-filling graphic designs in data visualization, can be better understood by the concept in data structures about development of tree maps the applications include: problems related to clustering, classification and regression trees, visualization of network, and many more
The extensive practice of descriptions in a variety of applications includes medicinal systems, classified armed records, enterprises and luggage compartment systems. Safekeeping is the foremost matter despite the fact that transmitting metaphors from first to last internet wherever hacking of top secret information might catch respite. Convey of imagery seeing as solitary position towards a different might infect the likeness in sequence, consequently novel tactic be resulting just before stop clatter as well as to get better the undisclosed illustration almost lossless.
Lately a variety of approaches came addicted to survival used for locked illustration broadcasting somewhere clatter issue has been not taken addicted to explanation and useful merely taking place RGB images. Consequently, a novel algorithm be intended toward condense the quantity of sound satisfied in the assortment representation and comprehensive to a different color model which be named like YUV, wherever authentic instant imagery be frequently exaggerated in AWGN.
Dissimilar conversion techniques have been implemented for reducing the noise property which comprises discrete wavelet transform (DWT), Stock well transform (S-Transform) and discrete curvelet transform (DCT). The planned performance takes RGB metaphors while key as well as convert them intense on YUV color model, resting on which the entire process has performed. The secret image, which be to exist transmitted in the direction of the headset, as well as cover images be segmenting into strips and blocks in that order of similar volume. The clandestine strips be fixed into cover up blocks along among color transformation be useful scheduled cover descriptions such to facilitate they appear alike to the wrap blocks.
The color altered ideals are Huffman encoded and stored in Huffman chart [1] which is inserted into the ensuing picture near appearance a mosaic image. It appears same while the cover image present next to dipping the hackers awareness for the period of communication. Whereas spread, the assortment representation (mosaic image) could be despoiled by noise and the noisy mosaic image be transferred the same as the input to the recipient.
The color altered standards be Huffman decoded and strips are extracted toward pick up strident secret image on or after the noisy mosaic image. The quantity of sound which is there in noisy secret image has been reduced by using a variety of transforms similar to wavelet, curvelet and S-Transforms to get back the YUV undisclosed image, which rotate into RGB color replica. In latest years, emergence of a lot of RDH techniques took place. Fridrichet al [2] constructed all-purpose agenda in favor of RDH by means of earliest method of compressing losslessly all the extracted compressible feature of original cover; this is done to save extra space in favor of embedding supporting information.
On the basis of difference expansion (DE)[3] numeral of well-liked technique has been used in which dissimilarity of the pixel excellence of both illustration be expanded as for example it strength of character be multiplied with 2 and therefore the least significant bits(LSBs) of the differences be all zeros and be able to be used for embedding the mails along with an extra individual be RDH for histogram shift(HS)[4] .This RDH approach is superior intended for histogram shifting .In which space has been saved in favor of embedding the information through the shifting the bins of histogram of gray scale .The state-of-art methods frequently pooled DE or else HS to residuals of the image, e.g., the predicted errors, to realize improved presentation.1.2 Mosaic Image
Mosaic Image is an ingenious effort somewhere a quantity of slighter imagery be resourcefully united as well as it be able to said to gathered in a single block designed for manufacturing the bigger representation. Every structural portion descriptions or floor covering have its own discrete and suggestive essence although the observation of Images from the far distance it’s seems like a fastidious Mosaic .This employment nearby the proposed and achievement of mosaic images.
A workstation software system to facilitate and generates these representation mosaics routinely. For controlling the mosaic images generation process different parameters has been used inside the system [5]. All parameters affect the largely assortment excellence and also affecting the processing time in its own unique way. A comprehensive study has been performed to calculate each factor independently. Furthermore this work proposed two narrative behaviors by which to appraise the superiority of a mosaic image in a quantify manner.
There are different method has been used where one method is used for perceptual color accuracy of the mosaic replica and another one is used for concentrates on border duplication [6]. Both are measuring unique visual features present in a mosaic Images. For minimizing eminence and due to the intrinsic properties of a mosaic images to facilitate them visually attractive.
Many instances show that a number of images are required to be acquire and then mosaiced to generate a more large and composite image. As far as, epi-fluorescence microscopy is concerned, one cannot characterize a complete tissue section of dimensions in single image of several millimeters, at high resolution. This is so because resolution necessary for a larger image to be viewed is high but such a high resolution cannot be afforded by a low power objective. If this high resolution is also achieved then also we will not be able to do so. Assembling the composite image from several images acquired at high magnification is the only obvious and practical solution to this.
1.3 Different Mosaic Techniques
1.3.1 Jigsaw Image Mosaic
Figure1.3: (a) Original image (b) Jigsaw Mosaic Image
In this technique image strip of subjective shape has been used to arrange the final depiction which are shown in figure 1.3. A jigsaw image mosaic is a type of puzzle representation. During this mosaics image strips of subjective character are used to create the final arbitrarily-shaped depiction called Jigsaw Mosaic Images.
1.3.2 Mosaic Images via Voronoi Diagram
A Voronoi diagram is an arithmetical arrangement to facilitate closeness information with reference to a set of points or substance. Particular a set of sites or items, the plane has been partitioned through conveying to each point to its nearest site. The points, whose adjacent site is not distinctive, structured the Voronoi diagram. There are two processes in the Method of Voronoi diagram. In the primary process, to remove the error between the original and the resulting image, repeated generation of the mosaic image is done by generation of the superlative Voronoi diagram.
Fig.1.4: (a) Original flower image (b) Final flower Image
The second procedure allows the consumer to add a variety of effect toward the assortment representation shaped by the primary step. The second procedure is planned in accordance with our surveillance of stained glass windows recognized that stained glass is one of the appliance that employ mosaic images. We can see the voronoi diagram representation in figure 1.4 where original image has been converted in puzzle form (voronoi form).
1.4 Secrete Fragment Visible Mosaic Images
In this transform a secret image and target images are converted into a significant mosaic image with the similar size and looking similar to a preselected target image. The alteration procedure is restricted via a secret key, and only through the key a person can recover the secret image nearly lossless from the mosaic image. A novel variety of computer art image called secret-fragment visible mosaic image has been projected, [1]-[2] which will be shaped through composing a small fragments of a given input image to become a target image in a mosaic form.
These properties hide the imagery and stay it top secret. In the direction of creating a mosaic image of this variety as of a given secret color image, the one color scale be changed into a new color scale, based on which a novel representation selecting as of a database as a target image is the greater part similar to the given secret image. From the figure 1.5 we can see secret image is first separated into rectangular formed fragments, called tile images, which are fixed into a target image.
Fig.1.5:(a) Secrete Image (b) Target Image (c) Secrete fragment Visible
Mosaic Image
1.5 YUV Color Model
YUV be a color space [5] previously used in the NTSC TV set. For in-phase, I is used whereas for quadrature, Q is used, referring to the mechanism used in quadrature amplitude modulation. YUV color space is now used by PAL and other system like NTSC. The Y constituents represent the luma information, and are the single constituent used by black-and-white television receiver. I and Q characterize the chrominance information. In YUV, the U and V machinery be able to think of seeing that X and Y coordinate inside the color space.
I and Q are able to take into consideration of as a second pair of axis lying on the same graph; rotated 33°; as a result IQ and UV represents dissimilar harmonize system on the similar plane. The YIQ scheme is planned to obtain benefit of human color-response individuality. The eye is more perceptive to change in orange-blue (I) range than in the purple-green range (Q) — so less bandwidth is necessary intended for Q and I. Broadcast NTSC confines I to 1.3 MHz and Q to 0.4 MHz I and Q are frequency interleaved interested in the 4 MHz Y signal, which keeps the bandwidth of taken as a whole signal downward near 4.2 MHz
In YUV system, from the time when U and V both contain information in the orange-blue range, both mechanisms have to go on specified the same amount of bandwidth as I to realize alike color commitment. Very few television sets perform true I and Q decoding, due to the high costs of such a realization. The Rockwell Modular Digital Radio (MDR) is one, which inside 1997 might activate in frame-at-a-time mode with a PC or in real-time with the Fast IQ Processor (FIQP).
YUV exist a color space usually using for measuring a color likeness channel. YUV encrypt a color image or video fastidious creature judgment into enlightenment, allowing draw round bandwidth used for chrominance machinery, in this way normally enabling propagation errors or density artifacts on the way to survive further competently enclosed during the individual surveillance, by means of an express RGB-representation. Complementary color spaces comprise correspondingly property, with the principal intention to put into practice or examine properties of YUV could be there intended for interfacing by means of analog or digital box or pictorial tools so as to conform to assured YUV standards.
The YUV color space has been used by the PAL (Phase Alternation Line), NTSC (National Television System Committee), and SECAM (Sequential Couleur Avec Memoires or Sequential Color with Memory) combined color video principles. The black and white system used only luma (Y) information; color information (U and V) was added in such a way that a black-and-white receiver would still display a normal black-and-white picture. Color receivers decoded the additional color information to display a color picture. RGB and YUV can be converted using the following algebraic equations:
Y = 0.299R ́ + 0.587G ́ + 0.114B ……………………… (1)
U = – 0.147R ́ – 0.289G ́ + 0.436B ́ = 0.492 (B ́ – Y) .……………………… (2)
V = 0.615R ́ – 0.515G ́ – 0.100B ́ = 0.877 (R ́ – Y) ….…….……………… (3)
‘OR’
R = Y + 1.140 ..….………………….. (4)
G = Y – 0.395U – 0.58V .….…………………… (5)
B = Y + 2.032U …….……………………(6)
For digital R G B values with a range of 0–255, ranges of Y, U and V are
• Y 0–255,
• U 0 to ±112, and
• V 0 to ±157.
To implement the above equations in a practical NTSC or PAL digital encoder or decoder, scaling of the equations are usually done. This is done to simplify the implementation process. Note that for digital data, 8-bit YUV and R ́G ́B ́ data should be saturated at the 0 and 255 levels to avoid underflow and overflow wrap-around problems. If the full range of (B ́ – Y) and (R ́ – Y) had been used, the composite NTSC and PAL levels would have exceeded what the (then cur-rent) black-and-white television transmitters and receivers were capable of supporting.
Experimentation determined that modulated subcarrier excursions of 20% of the luma (Y) signal excursion could be permitted above white and below black. The scaling factors were then selected so that the maximum level of 75% amplitude, 100% saturation yellow and cyan color bars would be at the white level (100 IRE).
1.6 Applications
NTSC (National Television System Committee):
YUV color representation is used for TV broadcast intended for historical reason. This scheme provisions aluma value by two chroma and chrominance values, resultant just about to the amount of blue and red in the color.
• PAL or (Phase Alternating Line):
Phase Alternation Line (PAL) covers most of the parts of Europe and is basically a system to encode colours in an analogue television. Just like the other TV standards like American National Television Systems Committee (NTSC) and the French Sequential Couleur avec Memoire (SECAM), PAL is also one of the three.
• SECAM or (Sequential couleur a amemoire, French for “sequential color with memory):
Most of the Asian Countries uses NTSC as their standard of TV including Japan. Most of the Western countries and Australia uses PAL or phase alternating line format whose bandwidth is quite large which gives a better picture quality. Another format popularly known sequential couleur avec memoire (sequential color with memory) or CAM is getting used in Eastern Europe and France.
1.7 Advantages
YUV is common in image and video compression (e.g. JPEG and MPEG). Image/video codec’s are YUV is so they can reduce the resolution of the U and V channels while keeping Y at full resolution, because luminance is more important than color. If you can reduce the resolution of U and V in a way that\’s compatible with convolution nets, your net should be half the size and therefore twice as fast, without much loss of accuracy.
• Human perception of colours can be better understood by the YUV color model than the computer graphics hardware RGB model
• It gives more Accuracy than RGB model.
Takes advantage of similarities within an image.
CHAPTER 2
LITERATURE SURVEY
In the field of Mosaic image creation different speedup techniques have been recommended. Main aim behind all of them is to surmount its main disadvantage i.e. Error rate and all others like SNR value, PSNR value. So literature survey and depth study is made of the methodology recommended by the various researchers and their presentation and results. These are arranged year wise in the following section.
1. C. Senthilkumar, “Color and Multispectral Image Compression using Enhanced Block Truncation Coding [E-BTC] Scheme”,
978-1-4799-5341-7/16/$31.00 ©2016 IEEE,
The main focus of this paper is on the problem of RGB color space identification from a single image in 2016. In photography, to produce an image, the use of RGB color spaces was most common. The issues as for shading space identification, for example, achieving steady printing or showing quality on screen, the main topic of concern should be gadgets and programming applications and aversion of mixed media unapproved use (appeared or printed by other gadget by means of array mapping). To concentrate shading space data, all the current procedures are dependent on Exchangeable Image File Format (EXIF). This paper utilizes a two-dimensional non-causal backward model to investigate the picture demosaicing properties. This is done mainly to concentrate discriminative elements and prepare the elements on SVM classifier for picture shading space recognition without being EXIF dependent. In our analysis, pictures in three diverse shading spaces (sRGB, adobe RGB and genius Photo RGB) are created for shading space identification undertaking. The trial comes about demonstrate that the proposed system has a decent execution on picture shading space identification.
2. Lizhi Wang, ZhiweiXiong, Guangming Shi, Wenjun Zeng and Feng Wu, “Compressive Hyperspectral Imaging with Complementary RGB Measurements”,
IEEE VCIP 2016, Nov. 27 – 30, 2016, Chengdu, China,
In this paper, to capture high quality images, a new camera which is hybrid in nature is designed for CASSI. Main advantage of this camera is that it maintains the snapshot advantage. Specifically, in this paper, along with the CASSI system, another complementary camera has been employed in 2016. We can get a spectral clue that is reliable in nature from this recorded RGB image. A hyperspectral image which is highly Fidel can be retrieved when the coded information of CASSI branch is combined with the uncoded information of RGB branch. Furthermore, to achieve a better performance, at preprocessing stage, the process of demosaicing can be performed on any raw RGB image. Theoretical analysis of the applied algorithm and its simulation resulted in improved accuracy as compared to the state-of-the-arts.
3. Huixuan Tang, Xiaopeng Zhang, ShaojieZhuo, Feng Chen, Kiriakos N. Kutulakos1 Liang Shen, “High Resolution Photography with an RGB-Infrared Camera”,
978-1-4799-8667-5/15/$31.00 ©2015 IEEE,
In the near infrared band, by combining the type fourth filter with high transmittance, the RGB mosaic can be extended. In infrared photography, this extension proved to be a more convenient solution. Unfortunately, in RGB-IR sensors, it is not possible to implement and apply conventional demosaicing algorithms for the following two reasons in 2015. First reason is, the refractive indices to different band are different, so the RGB and near-infrared image are focused differently. Second reason for the same is, introduction of crosstalk due to manufacturing constraints between RGB and IR channels. This paper has proposed a novel image formation model for RGB-IR cameras for easy calibration, and to address the problems like —channel deblurring, channel separation and pixel demosaicing, an efficient algorithm has been proposed —using quadratic image regularizers.
4. Ya-Lin Lee and Wen-Hsiang Tsai, “A new Secure Image Transmission Technique via Secret Fragment-Visible Mosaic Images by Nearly Reversible Color Transformations”
IEEE Transaction on Circuits and Systems for video technology volume. 24, no.4 April 2014,
It has proposed a new system in 2014 for increase the image transmission security. In this algorithm the secret image is converted into a significantly same sized mosaic image that looked like a target image that was preselected. Secret key controls alteration procedure and that secret image is only get well by that key exclusive of any loss from mosaic image Lai and Tsai unmitigated the planned method, in which a novel type of computer art image, called secret-fragment-visible mosaic image, was introduced. The mosaic image is the output of reorganization of the fragments of a secret image in camouflage of one more image called the target image preselected from information.
5. Anuprita U. Mande and Manish N. Tibdewal, “Parameter Evaluation and Review of Various Error-Diffusion Half toning algorithms used in Color Visual Cryptography”,
International Journal of Engineering and Innovative Technology (IJEIT) Volume 2, Issue 8, February 2013,
They have exhibited a strategy in 2013 for information covering up utilized as a part of \”shading video cryptography .They presented a mistake dissemination method for creating halftone offers which seems to be more wonderful to human eyes. From the survey of Color visual cryptography plans\”, it is seen that partially half conditioning of pictures is accomplished by different techniques in various plans. The paper took an audit of each of these techniques. In the meantime they have analyzed each of these techniques and received the one which gave them the best outcome as for shading visual cryptography.
6. JagdeepVerma, Dr. VineetaKhemchandani, “A Visual Cryptographic Technique to Secure Image Shares”,
International Journal of Engineering Research and Application (IJERA) Vol. 2, Issue 1, Jan-Feb 2012,
They have proposed a scheme that added the merits of the two i.e. visual cryptography as well as Invisible and Blind watermarking techniques in 2012, where they have generated the secretly shared images using basic visual cryptography model and then watermarking of these secret shares into some host image using invisible and blind watermarking is done. The process of decryption is done by stacking of the shares after the secret shares have been extracted by a simple “watermark extraction method”. The proposed watermarking schemes do not need the original image or any of its characteristics for the extraction of watermark, and hence the proposed scheme is shaded (blind).
7. J. Lai and W. H. Tsai, “Secret-fragment-visible mosaic image—A new computer art and its application to information hiding”,
IEEE Transaction InformationForens, Secure. vol. 6, no. 3, pp. 936–945, Sep. 2011,
They have proposed a special technique of hiding any information in 2011. This consisted of a secret image which is then divided into small fragments (tile images) of rectangular form and then to create a mosaic image they are fix to its next target image selected from a database. Secret key selects arbitrarily some blocks of mosaic images to implant the information of tile image. Without the key, recovery of secret information cannot be done by the hacker. The key can renovate the secret image by retrieving the entrenched information.
8. AtefMermoul, “An Iterative Speech Encryption Scheme Based on Subspace Technique”,
IEEE Transaction son Systems, Signal Processing and their Applications 2011,
An encryption scheme based on iterative speech has been developed based on subspace. Blind source separation (BSS)-based encryption schemes have been built up using the intractability of the under determined BSS problem. The author of this paper has designed and developed a scheme based on novel encryption that is iterative and also on the idea of subspace technique, by the nonlinear functions and the key signals. To decrypt any signal, only that part those parameters of secret key is required that were used in the process of encryption. If there is no plain-text fed in the input then there are no contents given by this technique.
9. Mohammad Reza Keyvanpour, FamooshMerrikh-Bayat, “A New Encryption Method For Secure Embedding In Image Watermarking”,
IEEE Transactions on Advanced Computer Theory and Engineering pp. 403-407, 2011,
To improve the security of data, the idea of watermarking based on coding and applying chaos function on the image, has been embedded in the process of encryption in this paper. In this technique, to get a good and secure transmission, the image pixels and their positions were rearranged using the Arnold\’s Cat Map method. Also, to get the feature of self similarity, division of the chaotic images into a range of blocks and domain of blocks is done. The process promotes a set of contractive transformations, for approximating the value of every block of the image to the larger block. This technique resulted in proving Chaos Fractal Coding algorithm to be best suited to get a good range of capacity and better security.
10. Suhaila O. Sharif, L.I. Kuncheva, S.P. Mansoor “Classifying Encryption Algorithms Using Pattern Recognition Techniques”
IEEE Transactions pp. 1168-1172, 2010,
To classify the algorithms for encryption, to make them in accordance with the Pattern Recognition method, a manuscript was framed jointly by the two. The discussion took the focus of the authors towards the limitations of the existing algorithms which are used for encryption scheme and for generating the keys for encryption process. In the process of encryption, for the identification of the block ciphers, here in this paper, they have used the pattern recognition method. The block cipher algorithms like AES, DES, IDEA, and RC were used to identify the good classification technique. Consequently, the performance of RoFo (Rotaion Forest) classifier has proved to be better than all the others and also it gives good classification accuracy.
11. Zhu Yuxi, Ruchun Cui, “Applied Study Based on OMAP Digital Fingerprint Encryption Method”,
IEEE Transactions pp. 1168-1172, 2010,
A newly invented technique of hiding the data has been proposed in this paper in 2006. This technique is popularly known as reversible data embedding technique. This is a technique in which, embedding a large amount of data (5–80 kb) can be easily done for a 512 (512 8 grayscale image). The graph plotted between the PSNR of the marked image versus the original image is guaranteed to be higher than 48 dB which is assumed to be a large percentage of visual quality for all natural images. For embedding any data, the use of zero or the smallest point of the histogram is done and also modifies the grayscale values of the pixels slightly. All types of images are based on this technique.
12. Huang, Jing, Zheng Zhen-zhuc, “A Method for Secure Real-Time Image Transmission Based on Optical Encryption”
International conference on the Intelligent Signal Processing and Communication Systems, 2010,
A scalable encryption method has been introduced in this paper. This method gives a good compatibility in the backward direction with the JPEG2000 format Images. In this technique, a multilevel encryption method is used that has decreased the complexity based on computation, in the encryption process. The paper has used JPEG2000 decoder at the decoding end where the encrypted images can be decoded properly and after the process of encryption, they have saved some parameters of JPEG 2000. To increase the speed or to fasten the process, this method has helped in controlling the duration or the time taken for the encryption by selective encryption algorithms.
13. B.V. Rama Devi et al., “A Novel Encryption Method for the Secure Transmission of Images”,
International Journal on Computer Science and Engineering, Vol. 02, No. 09, pp.2801-804, 2010,
Using key hopping method, this paper has proposed a technique for encrypting an image with enhanced Multiple Huffman Table (MHT). The existing Multiple Huffman Table (MHT) has shown good desirable properties but it was highly vulnerable to the chosen plaintext attack (CPA). But in this technique where the method has been enhanced, all the exiting limitations were eliminated to a great extent. As the result shown that the algorithm is secure for the chosen plaintext attack and proved mathematically by the key hopping method.
CHAPTER 3
IMAGE PROCESSING
3.1 Introduction
Image processing is a type of signal processing method intended for which the input is an image, such as a taken pictures or video structure and their yield might be both an image and a set of individuality or parameters connected to the depiction. Most image-processing techniques absorb treating the similarity as a two-dimensional signal and applying characteristic signal processing techniques just before it. Image processing refers to allowance of a 2D picture via computer. The technique used to improve raw image established from sensors sited on satellites, space probes and aircrafts or pictures in use in normal day-to-day life for different applications is Image Processing. Image processing system necessitates that the images be accessible in digitized form, that is, arrays of restricted length binary words. For digitization, the given Image is sampled on a discrete framework and each sample or pixel be quantized using a limited number of bits.
The digitized image has been processed by a computer. During the last few decades various techniques by different organizations have been developed for image processing. Most of the techniques are improved for enhancing imagery obtained from unmanned spacecrafts, space probes and military reconnaissance flights. Image Processing systems are fetching popular due to easy accessibility of authoritative personnel computers, a big size reminiscence devices, graphics software’s etc. Image Processing is used in different application such as:
• Remote Sensing
• Medical Imaging
• Non-destructive Evaluation
The common stepladder in image processing is image scanning, storing, enhancing and interpretation.
3.2 Methods of Image Processing
There are mainly two methods available in Image Processing.
3.2.1 Analog Image Processing
In this type of Processing, the image is altered with the help of some electrical means. Image obtained from and on television screen is the most common example of Analog Image Processing. The television signals which act as voltage levels have varied amplitude so that they can represent brightness through the image. Then appearance of the displayed image is altered by electrically varying signal. The TV sets which control the brightness and contrast, also regulates the amplitude and reference of the video signal, brightening, shadowing and fluctuation of the brightness in an image being displayed.
3.2.2 Digital Image Processing
For image processing, digital computers are used. In this process, an image is converted to digital form with the help of a scanner digitizer. Later it is taken into procedure. It can be defined as the subjecting numerical representations of objects to a series of operations in order to obtain a desired result. It begins as an image and then creates its modified form. When a digital computer processes a two-dimensional image, it is known as digital image processing. On the other hand, Digital Image Processing is used to digitize any two-dimensional data.
3.3 Image
What is an image? It is a function f(x, y) of two-dimensional light intensity. Here x and y indicate the co-ordinates and f, at any point (x, y), is proportional to, at that point, the gray levels or the intensity of the image. In this way, an image f(x, y), discretized together in spatial co-ordinates, will be a digital image. The image principles of pixels, then, can be termed as the principal of such a digital 11 array. In this way, an image can be defined basically as a two-dimensional signal by an individual signal.
In spite of the imagery, the signals are frequently in the analog form. However, they are to change from analog form to digital form to give out the storage and communication with processor application. A finite number of bits, representing a collection of authentic number is known as an image. Also, it is supposed to facilitate a digital image which is a 2-D collection of pixels. Images from the significant part of data, particularly in remote sensing, biomedical and video conferencing applications. Digital image processing method has some principal advantages as: repeatability, versatility and preservation of the original data precision.
3.3.1 Digital Image
The way of transforming pictures, text or sound from the analog media to the electronic media is known as digitization. We can keep these electronic data safe and organized. It can be regained or restored through some electronic device in any perceptible substitute of the primary work of those digital assets. Pictures, words and animations are predominated. Through photography, scanning, attachment or web downloading, digital images are created. These images may basically differ from their analog parts as slides or prints as they have no meaning without any software or a hardware that are used to translate or render them as an image. These digital images cannot be found or identified except the terms (Metadata) assigned to them.
To approach these digital forms of visual materials digitization helps a lot in “Random Filling”. Not any problem or worry is faced in maintaining and storing the images in their physical form.
Figure3.1: Digital Image
• Digital Images are attained in the form of pixel arranged in series
• The Pixel of each image carries a numerical value or Digital number
• Colors and shades of brightness are assigned to each digital number
3.3.2 Tile Image
Tiling process is applied for those images which are too large in shape or size. As for the picture of a satellite, which can be read only in gigabyte, can never be fitted into the memory or display. It is impossible to read it as a single object in an ordinary computer. Therefore it will have to be shown in segments using smaller tiles.
Figure3.2: Tile Image
The image can be created initially without any data when tiling is applied to it. The pixels of the image can be loaded only when the sections of the tile are viewed through panning. An image pyramid can also be created for the large images which will be supportive to the level of detail (LOD). Whenever we zoom in or zoom out an image, the resolution will be changed. A full view of that large image will quickly be provided. While zooming out, the images which are smaller in shape will be displayed. And we can see a full view of a large image.
Any particular area is chosen and is zoomed in. while doing so, the layers of the detailed image will be loaded progressively. Further the image will be aware of the required level of detail and the communication will be started with the application after requesting the visibility information of the tile.
3.3.3 Cover Image
A cover image is a kind of covering used to hide something behind. It is used to match the background colour perfectly with the image to conceal something.
3.3.4 Pixel Form
If the image is to be covered, the knowledge of Geometry is necessary for a digitizer or camera as while showing the pixels in figures, they appear as a square. The figure can help to define the features connected to a camera or a digitizer and their effect on the pixel.
Figure 3.3: Pixel Representation
3.4 Resolution
Figure 3.4: Picture Resolution in Different Sizes
Computer memory included the pixels, although the regions of finite area are the place where they are derived from, they can be considered as Mathematical points having no physical extent. While showing, the space lying between the points must be filled in. Normally it occurs as result of the finite spot size of a Cathode-ray tube (CRT). The CRT spot which has a brightness profile is Gaussian in some extent and the spots can be resolved on the display, it depends on the system quality. The display system of the resolution; 72 spots per inch (28.3 spots per cm) are to be obtained straightforwardly. The number further corresponds to the standard printing conventions. In case printing is not taken into consideration, higher resolution of about 30 spots per cm are attainable.
• Noise
Noise is any disturbing and unwanted signal that comes along with the desired content and consequently interferes in the process of transmission that takes place between the sender and the receiver. Images gathered and obtained on cameras and other modern sensors may contain this unwanted signal through noise sources. Noise is random in nature as it can be treated as opposite to deterministic distortions such as shading or lack of focus.
3.5 Image Formats
• GIF
The Graphics Interchange Format (.GIF) uses a color indexing format alike to that optionally provided by BMPs. A being GIF is competent of representing near 256 individual colors, so adaptation to this format will need a loss of information if the novel image contains additional colors than that. The data is added compacted with lossless LZW encoding.
• PNG
The Portable Network Graphics format (.PNG) was at first industrialized as an different to GIF, owing to bald-faced issues which have since expired. PNG has become fairly dissimilar from GIF, however, and have mostly displaced it for handling. PNG’s main facial appearances are clearness and a lossless firmness scheme that works fine on authentic images.
• JPEG
The Joint Photographic Experts Group came up with the now ever-present JPG/.JPEG standard. The most significant item to be responsive of about JPEG is that it is a lossy compression format. It wires a number of levels of firmness, with growing levels of information loss that result in artifact every so often known as the jaggiest.
3.6 Image Processing Steps
The following steps are followed in image processing.
• Image acquisition: This technique is used acquire an image which is digital.
• Image pre-processing: This step is to improve the image in ways that increases the probability for success of the other processes.
• Image segmentation: The process by which division of the input image into its constituent parts or items takes place.
• Image representation: This technique is used to make the image suitable for further processing.
• Image description: This technique takes out all the important information or features so as to differentiate single type items from one or more.
• Image recognition: This technique assigns some special tag to an item. This tag is assigned on the basis of the information provided by its descriptors.