N. M. Nandhitha

Euclidean Distance Based Color Image Segmentation of Abnormality Detection from Pseudo Color Thermographs

514 Abstract—Infrared thermography, non-contact, noninvasive technique is widely accepted as a medical diagnostic tool. An IR camera captures heat variations from the skin and maps into thermographs. Thermographs are acquired for the whole body or the region of interest. Thermographs either gray scale or pseudo color are processed for abnormality detection and quantification. However temperature variations are not normally visible to naked eye. Hence it is necessary to develop and analyze the feature extraction algorithms for abnormality detection. This paper proposes Euclidean distance based color image segmentation algorithm for abnormality extraction. Pseudo color thermographs of arthritis, stress fracture, ankle injury and fracture are considered.

Performance Evaluation of Image Processing Algorithms for Automatic Detection and Quantification of Abnormality in Medical Thermograms

Infrared Imaging is a totally non-invasive, non-contact, medical imaging procedure for detecting and monitoring various diseases and physical injuries. It diagnoses abnormal areas in the body by measuring heat emitted from the skin surface and expressing the measurements into a thermal map called thermograms. Abnormalities manifest as hot spots in thermograms. Thermologist has to interpret thermograms, identify and quantify the abnormality. To overcome the subjectivity involved in human interpretation, it is desirable to develop an image-processing algorithm for automatic interpretation of thermograms. Two different algorithms namely conventional image processing algorithm and region growing algorithm are proposed. Conventional algorithm involves thermogram acquisition, enhancement, segmentation, morphological processing and quantitative characterization. Region growing algorithm involves selecting a seed pixel and appending the similar pixels. The paper also compares the performance of these algorithms in terms of parameter dependency, image specificity and time consumption.

Text independent speaker recognition system using Back Propagation Network with wavelet features

Automated speaker recognition system is extremely important in areas such as Forensic and Defense. Performance of an automated speaker recognition system is dependent on feature extraction and classification. As speech is a non-stationary signal and the information is present in low frequencies, it necessitates a non-stationary tool that performs multi resolution analysis. An exhaustive approach is carried out in this work to identify the wavelet that is best suited for feature extraction. Of the various wavelets, Discrete Meyer Wavelet provides higher inter-class variance and lesser intra- class variance. Sixteen features are extracted for wavelet co-efficients and a five layered Back Propagation Network is used for recognizing the speakers.

Abnormality Detection from Medical Thermographs in Human Using Euclidean Distance Based Color Image Segmentation

Infrared thermography is recently widely accepted as a medical diagnostic tool. Thermographs are acquired for the whole body or the region of interest. Thermographs are processed for abnormality detection and quantification. As temperature variations are not normally visible to naked eye it is necessary to develop and analyze the feature extraction algorithms for abnormality detection. This paper proposes Euclidean distance based color image segmentation algorithm for abnormality extraction. Arthritis and stress fracture thermographs are considered

Detection and Quantification of Tungsten Inclusion in Weld Thermographs for On-line Weld Monitoring by Region Growing and Morphological Image Processing Algorithms

Tungsten inert gas welding is the best suited welding technique for precision welding in atomic and aircraft industries. The most commonly occurring weld defect is tungsten inclusion, which is mainly due to high welding current. Monitoring and controlling weld current can avoid the defect. Developing an automated on-line welding system to correct the deviation in the welding current requires an effective image-processing algorithm to extract defect features from the sensor output image. Infrared thermography is the best-suited sensor for on-line weld monitoring. Conventional feature extraction algorithms are parameter dependent, image dependent and time consuming thereby making it unsuitable for on-line monitoring. This paper proposes region growing and morphological image processing algorithm to identify and quantify the weld defect from thermographs. Defect is quantified by major axis length, minor axis length and area. It takes 0.406 seconds in contrast to 3.484 seconds of the conventional algorithm in a Pentium IV system with a processor speed of 2.64 GHZ system.

Extraction and quantification techniques for abnormality detection from medical thermographs in human

Nowadays Infrared thermography is widely accepted as a medical diagnostic tool. Thermographs are acquired for the whole body or the region of interest. Thermographs are processed for abnormality detection and quantification. As temperature variations are not normally visible to naked eye it is necessary to develop and analyze the feature extraction algorithms for abnormality detection. This paper analyzes the performance of three feature extraction algorithms used for abnormality detection. The developed techniques are applied on Arthritis and stress fracture thermographs.

ART Network Based Text Independent Speaker Recognition System for Dynamically Growing Speech Database

Automated recognizing a speaker from the speech signals is the foremost application in forensics. Speaker recognition system involves two phases namely feature extraction and a classifier system. Features extracted from the speech signals are fed to an already trained classifier system that identifies the speaker. Major challenge occurs when the database is periodically updated which necessitates retaining the classifier with new set of exemplars includes the old and new datasets. As training the neural network is computationally intensive, Back Propagation system is not ideal for speaker recognition system (updation). Hence it necessitates an efficient speaker recognition system that doesn’t forget the old database but adjusts to the new set of data. In this paper an Adaptive Resonance Theory (ART) based speaker recognition system is proposed that is capable of functioning well even in the case of periodic updation.

Automated weld defect classification from ultrasonic signals using statistical moments on normal distribution curves of wavelet co-efficient

Ultrasonic Testing is a highly reliable Non-Destructive Testing Technique for weld defect characterization. Defects occur either high frequency components (Porosity, Sidewall crack) or as low frequency components (Root, Lack of Fusion, Lack of penetration, slag) in the UT signal. Manual interpretation of these signals is subjective in nature and is dependent on the expertise of the individual. Hence it is necessary to develop automated signal analysis system that classifies the defect. As defect classification is non-linear in nature, neural network based classification techniques are cited in literature. However neural network based techniques are computationally complex and has prediction error. Hence in this paper, an effective range based classification system using statistical moments is proposed. Performance of the proposed technique is measured in terms sensitivity and specificity.

PERFORMANCE EVALUATION OF HOT SPOT EXTRACTION AND QUANTIFICATION ALGORITHMS FOR ON-LINE WELD MONITORING FROM WELD THERMOGRAPHS

The quality of welded steel structures play an important role in determining the reliability of a building. Weld quality is affected by Incomplete Penetration, which is a most commonly occurring defect in welds. An automated adaptive welding system if developed can correct the deviation in the welding current online so as to adjust the depth of Penetration to provide defect free welds. This system requires an on-line weld-monitoring sensor, efficient image processing algorithm for defect identification and neurofuzzy control software for correlating the defect characteristics with deviations in physical parameters. Infrared Thermography is the best-suited sensor for on-line weld monitoring and continuous assessment of welds. Incomplete penetration affects the hot spot of the thermograph and hence hot spot quantification describes the defect effectively. Three different algorithms namely conventional algorithm, region-growing algorithm and Euclidean distance based color image segmentation algorithm are developed for hot spot quantification. The paper compares the effectiveness and suitability of these algorithms for on-line weld monitoring.

Study on the impact of CDFG on the design aspects of combinational logic circuits

CDFG plays the most important role in digital circuit design as it directly affects the area, speed and cost of the designed circuit. Hence in order to obtain an effective circuit, it is necessary to optimize CDFG. In this paper a detailed study on CDFGs for four different circuits is made for both pipelined and non-pipelined architecture. Performance is Measured in terms of number of circuit elements, delay and Balsa cost, it is found that the area complexity increases in the CDFGs of pipelined architectures. On the other hand, the CDFGs of non-pipelined architecture are computationally simpler. However latency decreases in the former when compared to the later.