Ch Amarnatha Sarma ¹, K. Murali Krishna ¹, K. Rasool Reddy ¹

Affiliations
1 ECE Department, GEC, Gudlavalleru, A.P, IN

Abstract

Expression detection is valuable as a non-invasive procedure of lie detection and behavior prediction. Nevertheless, these facial expressions may be problematic to realize to the untrained eye. There are a number of techniques are to be had to establish the facial expressions in that we implements facial features recognition strategies making use of Principal Component Analysis (PCA) and Wavelet. The facial features Experiments are performed making use of our possess database. The universally accepted three foremost emotions to be recognized are: shock, unhappy and happiness together with impartial. Firstly now we have tried to evaluate facial evaluation and then participate in the PCA and Euclidean distance established matching Classifier is used to notice and classify the facial expressions. All these experiments are implemented using MATLAB 2013a. Utilising PCA and Wavelets the got results has reached a attention cost of 91.Seventy eight%.

Keywords

Facial Expressions, Principal Component Analysis (PCA), Wavelet Transform.

Full Text

   

K. Rasool Reddy ¹, M. H. Sundeep Singh ¹, K. G. V. Nageswara Rao ¹, M. Kama Raju ¹

Affiliations
1 Department of ECE, Gudlavalleru Engineering College, Vijayawada – 521356, Andhra Pradesh, IN

Abstract

In present day’s security from unauthorized humans (intruders) and fire/flame is one of the essential concerns especially in industries, banks, airports, railway stations and other restricted areas. Therefore, we introduce an Electronic Security System (ESS). The ESS provides the attentive message from flame as well as intruders. The proposed system generally consists of 3 devices namely PIR sensors, MSP430 and GSM. The flames and intruder are recognized by PIR. These activities are feeding to HMM, which are synthesized by wavelet transform. When there is an action occurs within the sensor range, then signal from sensor is processed by biorthogonal wavelet transform and then feeding to HMMs. The detection of intruder and flame is updated according to the HMM producing the highest probability. After that the alert message is sent to authorize person using GSM.

Full Text