Accepted Papers
- An Adaptive Nonlocal Mr Image Denoising Based on Maximum Likelihood Estimation using Local Neighborhoods
Teenu Thomas V , Adi Shankara Institute of Engg and Technology , IndiaABSTRACT
MRI Denoising is the process of removing the noise elements from an input data. The major challenge present in this field is to denoise the MRI without removing the finer details in the input data. Estimation and denoising MR images is important for proper medical diagnosis. The principal source of noise in MRI is thermal in origin which is produced electrons. The source of noise in MRI is these randomly fluctuating currents in the sample (imaged object) and in the receiver coil. The estimation of noise level in the MRI is by local maximum likelihood estimation of the input image. After estimating the noise levels in the MRI image the denoising can be performed. In the thesis phase I the adaptive nonlocal and nonlocal denoising methods are compared. And adaptive nonlocal method performs efficiently than nonlocal denoising. In the thesis phase II the denoising can be improved by using the restricted local neighborhoods algorithm. In the existing system the restricted local neighborhoods uses a reference image created by using nonlocal maximum likelihood estimation method. In the proposed system the reference image creation algorithm is replaced by the adaptive nonlocal method. - Data hiding in compressed & Encrypted Images
Ancymol Anto, Adi Shankara Institute of Engg and Technology , IndiaABSTRACT
Data hiding represents a class of processes used to embed data, such as copyright information, into various forms of media such as image, video, audio or text with a minimum amount of perceivable degradation to the “host” signal. The safest method to keep messages transmitted through open channels from leaking out is to encrypt them into a meaningful content. With the explosive growth of internet and the fast communication techniques in recent years the security and the confidentiality of the sensitive data has become of prime and supreme importance and concern. To protect this data from unauthorized access and tampering various methods for data hiding like cryptography, hashing,authentication have beendeveloped and are in practice today.And this datas can be hided in any mediums like texts,audio,video,and images. Image data hiding is a covert communication method that uses an image as the cover to hide the truth from potential attackers that some secret message hidden in the image is being transported Images are the most powerful medium for data hiding because of the limitation of Human visual System(HVS). The data hiding can be done in two different domains.They are spatial and frequency domains.In the spatial domain the cover image is compressed for more area.So the compression can be done by the Block Truncation coding which is most efficient compression technique.Then the data is hided in the compressed images.The proposed system has four parts such as image encryption data embedding image decryption and data extraction.The image is encrypted by taking xor operation of the input image and the entered encryption key value.So to decrypt and obtainthe same image the same key have to be enterd.Otherwise a wrong image will be obtained.To do the data embedding the pixel grouping have to be done.Then the space for the data will be obtained.So the data can be embedded in it.then the decryption can be done by using the same key used for encryption. - Make a Difference in Wigner-Vile Distribution to Improve Feature Extraction in Electroencephalogram Signal to Recognize Epileptic Patterns
Moslem Sharifzadeh and Farshad Tajeri , Shiraz University , Iran
ABSTRACT
Importance of considering epilepsy is the number of people suffer from it. About one percent of world population at least experience epilepsy in their lifetime. In this paper electroencephalogram signals are classified to recognize epileptic signal. Features which is needed to classify are extracted in timefrequency plane of a database which is recorded by Andrzejak. Novelty of this work is to use a new kernel instead of instantaneous correlation which Winger-Vile used. Neural network is used to classify data, result of classifying compared with other works. Accuracy of this classifier is 96.25% for five sets. - Reversible Data hiding in images using Histogram shifting
Nivya Gopal and Deepika M P , Adi Shankara Institute of Engg and Technology, IndiaABSTRACT
The concept of reversible data hiding technique is based on steganography and relatedwith internet security. When it is desired to send the confidential/important/secure data over an insecure channel it is necessary to encrypt as well as compress the cover data and then embed the confidential/important/secure data into that cover data. For achieving this facility there are various data hiding techniques, compression techniques, encryption/decryption techniques available. Also it is important that the data hiding should be reversible in nature, should be suitable for encryption/decryption domain. The main issues regarding the data hiding are Distortion, Noise condition and uneven distribution of embedding capacity. In this paper, we propose a novel method by reserving room before encryption with a traditional RDH algorithm, and thus it is easy for the data hider to reversibly embed data in the encrypted image. The proposed method can achieve real reversibility, that is, data extraction and image recovery are free of any error. - Teeline Shorthand Character Recognition using Artificial Neural Network
Nirmala. M and Nandhini. V , Sona College of Technology , IndiaABSTRACT
This paper demonstrates the use of neural networks for developing a system that can recognize Teeline shorthand character and converted into respective English alphabets. In this paper we discuss implementing the architecture in real world character recognition. Each Teeline shorthand character is represented by binary values that are used as input to a simple feature extraction system, whose output is fed to our neural network system. The Artificial Neural Network is trained using Back Propagation algorithm.Which compromises Training, Calc- ulating Error, and Modifying Weights. A problem encountered with a 50 x 70 matrix is time-consuming training, which is conquered with Manual Compression. - Skin Colour Information and Morphology Based Face Detection Technique
Sharmila Kumari M, Akshay Kumar, Rohan Joe D’souza, Manjunath G K and Nishan Kotian , PA College of Engineering , IndiaABSTRACT
Locating and tracking human faces is a prerequisite for face recognition and/or facial expressions analysis, although it is often assumed that a normalized face image is available. In this paper, we propose a faster, yet efficient face detection approach based on mathematical morphology and skin colour information. We have devised some simple post-processing rules to eliminate non-face regions from face regions. Experimentation on our created database is conducted to reveal the performance of the proposed approach. - Vision based mobile Gas-Meter Reading
Mehdi Chouiten and Peter Schaeffer , WASSA , France
ABSTRACT
The constant increase of smartphones computation capabilities has allowed a growing number of applications. This combined with the improvements of sensors quality and to third generation (3G) and fourth generation (4G) telecommunication network coverage, made possible the development of robust and reliable computer vision applications exchanging significant amount of data. In the gas distribution industry, the consumption reporting is a very important issue. In France, the major gas provider (GDF Suez) plans to deploy 11 million smart meters within 2022. In the meantime, employees of GDF are periodically sent to manually collect data from customers.In this paper, we present a solution developed for GDF – Suez to solve this problem using mobile technologies and computer vision algorithms. - Malignancy detection in Cervical Cancer cells in Pap Smears Images using K-means clustering algorithm
Rimzhim Neog1, Lipi Mahanta2 and Usha Mary Sharma1 , 1Don Bosco College of Engineering and Technology , India and 2Institute of Advance Studies in Science and Technology , IndiaABSTRACT
Early detection of malignancy in human cells is very important for saving the life of a cancer patient. Sometimes the detection of cancerous cervical cells can be missed due to technical or human errors. In certain cases, the raw Pap smear images are distorted and highly affected by unwanted noises. These factors can hide and obscure the important cervical cell morphologies which increases the false diagnosis rate. In this paper, a novel image processing approach is proposed for the early detection of malignancy in Cervical Cancer cells in Microscopic Pap Smears Images using K-means clustering algorithm. - Overview of Speaker Recognition System
Rupali Pawar1, Rajesh Jalnekar2, Janardan Chitode2 and Rubeena Khan1 , 1MESCOE , India and 2Vishwakrma Institute of Technology, IndiaABSTRACT
Speaker Recognition is an important application of speech processing which enables us to recognize the speaker with the help of characteristics of their voice. The individual characteristics like pitch, fundamental frequency, formant frequency can be distinguishing components of the human speech signal. This paper gives an overview of various feature extraction techniques and Recognition techniques with their merits and demerits. It also discusses the pre-emphasis stage of the speaker recognition system. The various standard databases available for speaker recognition along with the criterion for their selection are also reviewed. The paper makes an attempt to discuss various tools and performance parameters of Automatic Speaker Recognition System. - Low Cost Real Time Robust Identification of Impulsive Signals
Gilles Ferone, Robin Biondi, Gareth Dys, Thibault Renard and Morgan Zysman , Paris School of Engineering , FranceABSTRACT
This paper describes an automated implementable system for impulsive signals detection and recognition. The system uses a Digital Signal Processing device for the detection and identification process. Here the system analyses the signals in real time in order to produce a particular response if needed. The system analyses the signals in real time in order to produce a specific output if needed. Detection is achieved through normalizing the inputs and comparing the read signals to a dynamic threshold and thus avoiding detections linked to loud or fluctuating environing noise. Identification is done through neuronal network algorithms. As a setup our system can receive signals to “learn” certain patterns. Through “learning” the system can recognize signals faster, inducing flexibility to new patterns similar to those known. Sound is captured through a simple jack input, and could be changed for an enhanced recording surface such as a wide-area recorder. Furthermore a communication module can be added to the apparatus to send alerts to another interface if needed. - A New Repetition Based Algorithm for denoising of Salt and Pepper noise.
Tejaswini Kar, Vinod Kumar, Priya Nayan and Richa Nayan , KIIT University , India
ABSTRACT
The main objective of the project is removal of salt and pepper noise which degrades the image quality. Salt and pepper noise involves noise value that may be maximum or minimum value of dynamic gray scale range of image respectively. The idea is to filter image having noise high density noise up to 92% and restore the original image. The algorithm makes use of methods based on median filtering in which corrupted pixels are replaced by median value or else if it is not effective then by the adjacent pixels. Run time, PSNR, SSIM and IEF were noted for different observations and compared. The major advantage of the proposed algorithm is the minimum run-time requirement and high PSNR value of the filtered image. The proposed algorithm is better even at high density noise and found to produce better PSNR and SSIM values - The Effect of Applying Gaussian Blur Filter on CAPTCHA’s Security
Ariyan Zarei , Shahid Beheshti University , IranABSTRACT
Providing security for webservers against unwanted and automated registrations has become a big concern. To prevent these kinds of false registrations many websites use CAPTCHAs. Among all kinds of CAPTCHAs OCR-Based or visual CAPTCHAs are very common. Actually visual CAPTCHA is an image containing a sequence of characters. So far most of visual CAPTCHAs, in order to resist against OCR programs, use some common implementations such as wrapping the characters, random placement and rotations of characters, etc. In this paper we applied Gaussian Blur filter, which is an image transformation, to visual CAPTCHAs to reduce their readability by OCR programs. We concluded that this technique made CAPTCHAs almost unreadable for OCR programs but, their readability by human users still remained high. - WhistleTronix – Frequency Detection and Processing of a Pressure Cooker Whistle Signal
Rashmi B, Riti Agarwal, Supriya Umashankar, Surabhi S, Yashaswini S and Veena S Murthy , BNM Institute of technology , IndiaABSTRACT
Electronics is all about making human lives simpler. This paper is one of the million efforts in that direction. Cooking is one of the most ancient parts of everyday activities and a pressure cooker has evolved to be an integral part of it. The main aim here is to eliminate confusions involved in the manual way of remembering the count. The heart of this design is the missing pulse detector, an application of 555 timers. This circuit detects the change in its input frequency and gives a non-oscillating output. Decision is made considering this as the criterion for further processing. - An Optimal Nonuniform Sampling
Kunal Sankhe and Garimella Rama Murthy , International Institute of Information Technology , IndiaABSTRACT
In this paper, a novel information theoretic based approach for nonuniform sampling is proposed. Using results from statistics, an “Optimal Nonuniform Sampling” is discussed.Problem for nonuniform allocation of samples has been formulated and solution based on Integer Programming has been proposed. - Iris Biometric System using a Hybrid Approach
Abhimanyu Sarin , BITS Pilani , UAEABSTRACT
Iris Recognition Systems are ocular- based biometric devices used primarily for security reasons. The complexity and the randomness of the Iris, amongst various other factors, ensure that this biometric system is inarguably an exact and reliable method of identification. The algorithm is responsible for automatic localization and segmentation of boundaries using circular Hough Transform, noise reductions, image enhancement and feature extraction across numerous distinct images present in the database. This paper delves into the various kinds of techniques required to approximate the pupillary and limbic boundaries of the enrolled iris image, captured using a suitable image acquisition device and perform feature extraction on the normalized iris image with the help of Haar Wavelets to encode the input data into a binary string format. These techniques were validated using images from the CASIA database, and various other procedures were also tried and tested. - Preprocessing on MRI of Brain
Ujwala Suryawanshi1, Santosh Chowhan2 and Uday Kulkarni3 , 1College of Computer Science & IT, India ,2S. R. T. M. University, India and 3SGGSIE&T , IndiaABSTRACT
This paper describes the preprocessing techniques for brain Magnetic Resonance Image (MRI) segmentation. Preprocessing is an important step in enabling accurate measurement of brain structures. Due to large amount of noise and non brain region the segmentation accuracy cannot be correctly obtained. Hence the preprocessing techniques are used, the image intensities are firstly standardized using the pixel histograms and Morphological operations. Morphological operations are applied to eliminate the non-brain regions or tissue, skull, dura from brain. The experimental result shows accuracy structure of Brain MR Images. - Fixed Range Block Partition and Classification for Fractal Image Compression of Satellite Imageries
Veenadevi.S.V. and A.G. Ananth , R.V.College of Engineering , India
ABSTRACT
The fixed range block segmentation and classification for fractal image compression has been carried out for Standard Lena and Satellite imageries. The encoding procedure consists of dividing the image into range blocks and domain blocks. Classification of sub image into three major classes and 24 different subclasses for every major class.There is total domain and range blocks are represented in 72 classes. Then, each range block is to search the corresponding domain block to find the best match. Affine transformation and entropy coding are applied to achieve fractal compression. The image is reconstructed using iterative functions and inverse transforms. The Matlab simulation for the standard Lena and Satellite imageries have been carried out for the fixed range block sizes of 4X4, 8X8 and 16X16. The results shows that for the fixed range block size 4X4 achieve higher Peak Signal to Noise Ratio (PSNR) ~ 28dB and Compression Ratio (CR) of ~ 3.7 for the Lena and Satellite imageries. The results for the fixed range block size 16X16 achieve higher CR of ~69 and PSNR of ~ 20dB for the Lena and Satellite imageries. This method is found to indicate higher PSNR, good CR and reduced encoding –decoding time. The results are presented and discussed in the paper. - Digital Image Restoration of Historical Devanagri Manuscripts
Nidhi Dubey , GLA University , IndiaABSTRACT
We live in a world where lots of manuscripts were written in earlier times. There are lots of historical manuscripts which are in deteriorated form. There is a need to restore these manuscripts in order to preserve our cultural heritage and ancient knowledge for future generation. Our focus is to restore the manuscripts that have been deteriorated because of age through some restoration techniques and binarization methods combining them with image processing methods applied on document images for text, background restoration on these deteriorated manuscript images - Teeth Classification in Dental Images using Support Vector Machine
Vijayakumari Pushparaj1, Banumathi Arumugam2 and Ulaganathan Gurunathan3 , 1Kamaraj College of Engg & Tech , India ,2Thiagarajar College of Engineering , India and 3Best Dental Science College , IndiaABSTRACT
In the present day world, individual identification with some reliable means is emerging as a significant state of affairs. Since teeth pattern is unique for individual human being, it can be treated as a suitable biometric means. It is playing the major role during mass disaster identification and individual identity. In order to ease the process of human identification using dental images, teeth classification is desired as an imperative process. This paper introduces teeth classification using linear and multi-class support vector machine. Teeth information can be acquired by either radiographic or photographic means. The algorithm is implemented by performing preprocessing initially, then teeth separation followed by feature extraction and classification. The accuracy of linear support vector machine yields 91% for radiographs and 95% for photographs in terms of number of teeth tested and correctly classified. Multi class support vector machine improves the performance of classification with inclusion of canine teeth in radiographs, achieved accuracy of 90.5%, which is comparable with the existing algorithms. - A 5.99 GHz Inductor-less Current Controlled Oscillator for High Speed Communications
Chakaravarty Rajagopal and Othman Sidek , University Science Malaysia , MalaysiaABSTRACT
This paper presents the design of five-stage current controlled inductor-less ring oscillator that were simulated in Silterra 0.18um CMOS Technology with oscillation frequencies up to 5.99 GHz. The design uses cross coupled MOS devices along with active inductor (thus inductor-less) and controlled by current source to aid in switching speed and to improve the noise parameters. The simulations show that the five-stage oscillator achieves frequency in the range of 3.78GHz to 5.99GHz. The simulated phase noise of the same design was -115.67 dBc/Hz at 1MHz offset with a center frequency of 5.99GHz