Ratnadeep R. Deshmukh

Indian Language Speech Database: A Review

is the most prominent and natural form of communication between humans. Human beings have long been motivated to create computer that can understand and talk like human. When the research tries to develop certain recognition system they require certain previously stored data i.e. database for respective recognition system. There are various speech databases available for European Language but very less for Indian Language. In this paper we discuss the various Speech Database developed in different Indian Languages for speech recognition system & Text to Speech System.

A Comparative Study of Feature Extraction Techniques for Speech Recognition System

The automatic recognition of speech means enabling a natural and easy mode of communication between human and machine. Speech processing has vast applications in voice dialing, telephone communication, call routing, domestic appliances control, Speech to Text conversion, Text to Speech conversion, lip synchronization, automation systems etc. Here we have discussed some mostly used feature extraction techniques like Mel frequency Cepstral Co-efficient (MFCC), Linear Predictive Coding (LPC) Analysis, Dynamic Time Wrapping (DTW), Relative Spectra Processing (RASTA) and Zero Crossings with Peak Amplitudes (ZCPA).Some parameters like RASTA and MFCC considers the nature of speech while it extracts the features, while LPC predicts the future features based on previous features.

A Review on Different Approaches for Speech Recognition System

This paper presents the basic idea of speech recognition, proposed types of speech recognition, issues in speech recognition, different useful approaches for feature extraction of the speech signal with its advantage and disadvantage and various pattern matching approaches for recognizing the speech of the different speaker. Now day’s research in speech recognition system is motivated for ASR system with a large vocabulary that supports speaker independent operations and continuous speech in different language.

Hybrid Feature Extraction Technique for Face Recognition

This paper presents novel technique for recognizing faces. The proposed method uses hybrid feature extraction techniques such as Chi square and entropy are combined together. Feed forward and self-organizing neural network are used for classification. We evaluate proposed method using FACE94 and ORL database and achieved better performance.

Analysis of Retina Recognition by Correlation and Covariance Matrix

We present an automated technique for person recognition based on retina of the human eye. In this paper we compare the performance of retina recognition by calculating correlation and covariance matrix of the retinal images. 20 images are used for the purpose of training and testing. Experimental results on DRIVE database show that these two methods are significantly better

Face recognition using fusion of PCA and LDA: Borda count approach

Face recognition has become one of the most successful applications in the field of image analysis and understanding. This paper presents a new approach for identity recognition using rank-level fusion of multiple face representations. In this paper, we propose face recognition based on fusion of two well-known appearance-based techniques, Principal Component Analysis and Linear Discriminant Analysis. Fusion is done at rank level using Borda count method. Our experimental work demonstrates significant improvement in recognition accuracy over individual face representations.

Biometric Template Protection with Fuzzy Vault and Fuzzy Commitment

Conventional security methods like password and ID card methods are now rapidly replacing by biometrics for identification of a person. Biometrics uses physiological or behavioral characteristics of a person. Usage of biometric raises critical privacy and security concerns that, due to the noisy nature of biometrics, cannot be addressed using standard cryptographic methods. The loss of an enrollment biometric to an attacker is a security hazard because it may allow the attacker to get an unauthorized access to the system. Biometric template can be stolen and intruder can get access of biometric system using fake input. Hence, it becomes essential to design biometric system with secure template or if the biometric template in an application is compromised, the biometric signal itself is not lost forever and a new biometric template can be issued. One way is to combine the biometrics and cryptography or use transformed data instead of original biometric template. But traditional cryptography methods are not useful in biometrics because of intra-class variation. Biometric cryptosystem can apply fuzzy vault, fuzzy commitment, helper data and secure sketch, whereas, cancelable biometrics uses distorting transforms, Bio-Hashing, and Bio-Encoding techniques. In this paper, biometric cryptosystem is presented with fuzzy vault and fuzzy commitment techniques for fingerprint recognition system.

Adaptive Apriori Algorithm for frequent itemset mining

Obtaining frequent itemsets from the dataset is one of the most promising area of data mining. The Apriori algorithm is one of the most important algorithm for obtaining frequent itemsets from the dataset. But the algorithm fails in terms of time required as well as number of database scans. Hence a new improved version of Apriori is proposed in this paper which is efficient in terms of time required as well as number of database scans than the Apriori algorithm. It is well known that the size of the database for defining candidates has great effect on running time and memory need. We presented experimental results, showing that the proposed algorithm always outperform Apriori. To evaluate the performance of the proposed algorithm, we have tested it on Turkey students database as well as a real time dataset.

Effect of Distance Measures on the Performance of Face Recognition Using Principal Component Analysis

We examined effect of distance measures on the performance of face recognition using Principal Component Analysis. We tested commonly used 4 distance measures: City block, Euclidean, Cosine and Mahalanobis. The study was done on 400 images from ORL face database. We achieved the best recognition performance using Cosine and City block distance measures. It was observed that fewer images need to be extracted for achieving 100% cumulative recognition using Cosine metrics than using any other distance measure. The performance of Mahalanobis distance measure was poor compared to other distance measures. Our results show the importance of using appropriate distance measure for face recognition.

Emotion Classification of Restaurant and Laptop Review Dataset: Semeval 2014 Task 4

The ―Sentiment Analysis‖ task focuses on the recognition and classification of emotions (positive, negative, conflict, neutral) in reviews for the aspect. In this paper we propose the system for recognizing and analyzing the sentiments using SVM for the restaurant and laptop review dataset. We compare the performance of the system with well-known KNN classifier. Index Terms Aspect, Sentiment Analysis, SVM, KNN