L. Padma Suresh

Image segmentation using seeded region growing

Image segmentation is the process of clustering pixels into salient image regions (i.e) regions corresponding to individual surfaces, objects or natural parts of objects. Image segmentation plays a vital role in image analysis and computer vision applications. Several general-purpose algorithms and techniques have been developed for image segmentation. Segmentation process should be stopped when region of interest is separated from the input image. Based on the application, region of interest may differ and hence none of the segmentation algorithm satisfies the global applications. Thus segmentation still remains a challenging area for researchers. This paper presents a comparison of some literature on color image segmentation based on region growing and merging algorithm. Finally an automatic seeded region growing algorithm is proposed for segmenting color images.

Literature survey on face and face expression recognition

Face Expression Recognition (FER) has become a very interesting and challenging area in computer vision field due to its wide application possibilities. Mental state Recognition, Human Computer Interaction, Human behavior understanding etc. are some of its applications. Because of its wide application possibilities Face expression recognition has attained a very crucial role in the area of facial image processing. In this paper some of the tailor made face expression Recognition algorithms are presented. This paper also gives a brief insight into the feature extraction method of these face expression recognition techniques. The features extraction technique plays a crucial role in the efficiency of these algorithms. In this paper a few Face Expression Recognition techniques like Patched Geodesic Texture Transform, Curvelet Feature Extraction, Bag of Words Method, Local Directional Number Pattern, Regional Registration Technique, Gradient Feature Matching etc. which are used to recognize the facial expression are presented.

Image segmentation based on genetic algorithm for region growth and region merging

Medical image segmentation is the most important process to assist in the visualization of the structure of importance in medical images. Malignant melanoma is the most frequent type of skin cancer but it is treatable, if diagnosed at an early stage. Dermoscopy is a non-invasive, diagnostic tool having great possibility in the early diagnosis of malignant melanoma, but their interpretation is time consuming. In this work, a new image segmentation algorithm, for the early diagnosis of the skin cancer, is proposed where the dermoscopic images are segmented using a threshold. This threshold selection is based on the Genetic Algorithm (GA) for region growth, followed by region merging procedure. The obtained segmented image is then compared with the ground truth image using various parameters such as False Positive Error (FPE), False Negative Error (FNE) Coefficient of similarity, spatial overlap. Genetic algorithm is a class of probabilistic optimization algorithms, powerful in finding optimal feature vectors. Identification of approximate global optimal region in Genetic Algorithm is a quick process. To merge regions with similar characteristics, we have used grey level and texture. The segmentation done through Genetic Algorithm is efficient when compared to the image segmentation done by conventional algorithms.

Image steganography using mod-4 embedding algorithm based on image contrast

In order to improve the capacity of the hidden secret data and to provide an imperceptible stego image quality, a new image steganography method based on image contrast is presented. A group of 2×2 blocks of non-overlapping spatially adjacent pixels is selected as the valid block for embedding the secret message. The modulo 4 arithmetic operation is further applied to all the valid blocks to embed a pair of binary bits using the shortest route modification scheme. Each secret message is also encrypted by RSA encryption algorithm to provide the system with more security. Data will be embedded inside the image using the pixels. Then the pixels of stego image can then be accessed back in order to retrieve back the hidden data inside the image. However, a secret key is needed by the receiver in order to retrieve back the data. This secret key is generated using the RSA decryption algorithm. By using the secret key to retrieve the data, it maintains privacy, confidentiality and accuracy of the data. The proposed method was tested on different gray scale images. From the experimental results, compared with the some well-known adaptive and non-adaptive steganography algorithms, the proposed method provides larger embedding capacity, while being less detectable by steganalysis methods.

A brief review on multi level inverter topologies

In this paper a brief review on different multilevel inverter topologies are discussed. Inverter is a power electronic device that converts DC power into AC power at desired output voltage and frequency. Multilevel Converters nowadays have become an interesting area in the field of industrial applications. Conventional power electronic converters are able to produce an output voltage that switches between two voltage levels only. Multilevel Inverter generates a desired output voltage from several DC voltage levels at its input. The input side voltage levels are usually obtained from renewable energy sources, capacitor voltage sources, fuel cells etc. The different multilevel inverter topologies are: Cascaded H-bridges converter, Diode clamped inverter, and Flying capacitor multilevel inverter. Multilevel inverters nowadays are used for medium voltage and high power applications. The different field of applications include its use as UPS, High voltage DC transmission, Variable Frequency Drives, in pumps, conveyors etc. The disadvantages of MLI are the need for isolated power supplies, design complexity and switching control circuits.

Dermoscopic image segmentation and classification using machine learning algorithms

Dermoscopy is the method of examining the skin lesions. It is especially used for diagnosing melanoma, a type of skin cancer. Image segmentation and classification are important tools to provide the information about the Dermoscopic images clinically in terms of its size and shape. Many algorithms were developed for classification and segmentation of Dermoscopic images. This work proposes the tasks of extracting, classifying and segmenting the Dermoscopic image using the machine learning algorithms. The algorithms such as Back Propagation network (BPN), Radial Basis Function Network (RBF) and Extreme Learning Machine (ELM) are used. The features are extracted from the Dermoscopic image and these features are used to train the classifiers. The trained networks are used for segmentation. The results are compared with the ground truth images and their performance is evaluated. The results proved that the ELM has better accuracy, faster training period and it provides better segmentation than the BPN and RBF neural networks.

Adaptive neuro-fuzzy inference system for classification of ECG signal

The heart is one of the important parts of any human being. The heart produces electrical signals thus electrical signals are normally called as Electrocardiogram (ECG) signal. The Electrocardiogram signal is used for identifying the heart problems. The objective of this work is to implement an ANFIS algorithm for (ECG) signals classification. In this work, the classification is done using the ANFIS associated with back propagation algorithm. The ANFIS model is combination of adaptive capabilities with neural network the qualitative approach of fuzzy logic. The feature selection process is done before classification. Four types of ECG beats are collected from the PhysioBank databases. These heart signals are classified by four ANFIS classifiers. The fifth ANFIS classifier is used to get an improved diagnostic accuracy in the ECGs.

A choice of FPGA design for three phase sinusoidal pulse width modulation

A FPGA based three phase sinusoidal pulse width modulator IC is designed. Pulse width modulated dc-ac converters have a broad range of applications in ac motor drives and ac power conditioning systems. The PWM strategy plays important roles in the reduction of harmonics and switching losses in these converters especially in three phase applications. Various modulation strategies, control schemes has been developed in recent years. In this work it is proposed to design FPGA/CPLD based ASIC solution for sinusoidal pulse width modulation generation with reconfigurable technology. It is also proposed to compare different architectures based on number of memories, multipliers and comparators for the generation of SPWM in order to optimize the speed and size of the SPWM generator IC. This SPWM generation is simulated by ModelSim SE 6.3f and synthesis by Xilinx Project Navigator tool. The proposed IC for SPWM generation will facilitate faster but compact solution for three phase voltage source inverter fed drives.

Design of FIS-Based Model for Emotional Speech Recognition

Human beings have emotions associated with their acts and speeches. The emotional expressions vary with moods and situations. Speech is an important medium through which people express their feelings. Prosodic, spectral, and other parameters of speech vary with the emotions. The ability to represent the emotional speech varies with the type of features chosen. In an attempt to recognize such an emotional content of speech, one of the spectral features (linear prediction coefficients), have been first tested by the fuzzy interference system. Next to it hybridization of LPC features with different prosodic features were compared with LPC features for recognition accuracy. Results show that the hybridization of features can classify emotions better with the FIS system.

Classification of ECG Signal Using Hybrid Feature Extraction and Neural Network Classifier

The heart is one of the crucial parts of a human being. The heart produces electrical signals and these cycles of electrical signals are normally called as cardiac cycles. The graphical recording of the cardiac cycles is known as Electrocardiogram (ECG) signal. The Electrocardiogram signal is used to diagnose the irregularity in heart beat which in turn can be used to identify heart problems. Automatic classification of ECG signals has applications in human-computer interaction, as well as in clinical application such as detection of key indicators of the onset of the certain illness. In this work, we have developed an algorithm to detect the five abnormal beat [2, 3] signals includes Left bundle branch block beat (LBBB), Right bundle branch block beat (RBBB), Premature Ventricular Contraction (PVC), Atrial Premature Beat (APB) and Nodal (junction) Premature Beat (NPB) along with the normal beat. In order to prepare an appropriate input vector for the neural classifier several pre processing stages have been applied. For efficient feature extraction we use hybrid feature extractor. The hybrid feature extraction is done in three steps, (i) Morphological based feature extraction (ii) Haar wavelet based feature extraction (iii) Tri-spectrum based feature extraction. The classification is done using Forward Feed Neural Network. Finally, the MIT-BIH [4] database is used to evaluate the proposed algorithm. The beat classification hybrid system (Hybrid + FFBN) based gives an accuracy is achieved 78 %, (Morp + FFBN) is achieved 62 %, (wavelet +FFBN) is achieved 65 %, (Spect + FFBN) is achieved 70 %, (Morp + Wavelet + FFBN) is achieved 62 %, (Morp + spect + FFBN) is achieved only 73 %.