DEVELOPMENT OF A DISCRETE-FIREFLY ALGORITHM BASED FEATURE SELECTION SCHEME FOR IMPROVED FACE RECOGNITION
Need help with a related project topic or New topic? Send Us Your Topic
DOWNLOAD THE COMPLETE PROJECT MATERIAL
DEVELOPMENT OF A DISCRETE-FIREFLY ALGORITHM BASED FEATURE SELECTION SCHEME FOR IMPROVED FACE RECOGNITION
Chapter One: Introduction 1.1 Background of Research
Researchers are interested in face recognition due to its wide range of applications, including public security, law enforcement, credit card verification, criminal identification, access control, human-computer interaction, digital libraries, and information security (Bakshi & Singhal, 2014). Face recognition compares a person’s biometrics to a database of pre-defined faces to authenticate their identification (Zhou et al., 2014).
According to Shivdas (2014), the face plays a significant role in identifying individuals throughout daily interactions. Face recognition outperforms traditional methods of identification, such as passwords and personal identification numbers
due to its accuracy and case sensitivity (Angle et al., 2005; Kaur & Singh, 2015). It is also non-contact, can be easily captured or videoed, and has a wide range of applications (Bakshi & Singhal, 2014).
According to Maini and Aggarwal (2009), the face is crucial in transmitting originality and emotion in public. Face recognition involves three stages: detection, feature extraction, and classification or recognition (Maini & Aggarwal, 2009).
Choosing the appropriate technique for each stage is crucial for improving recognition accuracy. Face identification is difficult by factors such as age, skin colour, gender, image quality, facial expressions, backdrop, and lighting conditions (Bakshi & Singhal, 2014).
Face detection aims to identify an object in a picture that closely resembles a human face (Saleh, 2009). This approach involves automatically recognising a face from a complicated background and applying a face recognition algorithm. Researchers utilise pre-processing throughout this stage (Agarwal & Bhanot, 2015).
Hemalatha and Govindan (2015) describe feature extraction as the process of extracting high-level information on individual patterns, such as eyes, nose, and mouth, to aid in recognition.
Choosing the right feature extraction strategy is crucial for optimal recognition performance (Saleh, 2009). Face extraction methods include discrete cosine transform (DCT) (Hemalatha Gayatri & Govindan, 2015; Jadon et al., 2015), gabor filter (Keche et al., 2015; Ruan et al., 2010), principal component analysis (PCA)
(Bakshi & Singhhal, 2014; Satone & Kharate, 2014; Sawalha & Doush, 2012), local binary pattern (LBP) (Babatunde et al., 2015), and discrete wavelet transform (DWT) (Kallianpur et al.).
During the classification or recognition stage, face samples are compared to known faces in the database (Richa & Josan, 2013). Several techniques have been published at this level, including support vector machine (SVM) (Satone & Kharate, 2014; Xu & Lee, 2014), Hidden Markov Model (HMM) (Jameel, 2015), and Nearest Neighbour Classifier (NNC) (Agarwal & Bhanot, 2015).
Shivdas (2014) and Bakshi and Singhal (2014) use Back Propagation Neural Network (BPNN) and self-organizing map (SOM) respectively.
Makiantan et al. (2012) describe the feature selection procedure, which involves determining the optimal subset to represent a given set. The feature selection problem is complex due of its combinatorial character.
The feature selection phase of face recognition aims to identify the main distinguishing traits between two or more faces, capturing differences in illumination, position, expression, or occlusion (Agarwal & Bhanot, 2015).
This leads to higher accuracy in databases. Excessive features can lead to overfitting of face data, resulting in poor system performance (Agarwal & Bhanot, 2015). Various optimisation techniques are used in feature selection, including particle swarm optimisation (PSO) (Hemalatha & Govindan, 2015; Ramadan & Abdel-Kader, 2009; Unler & Murat, 2010; Xue et al., 2014)
firefly algorithm (Agarwal & Bhanot, 2015), genetic algorithm (GA) (Boubenna & Lee, 2016; Satone & Kharate, 2014), and ant colony optimisation (ACO) (Babatunde et al., 2015; Kanan & Fa).
To improve recognition accuracy, this research proposes using a discrete firefly algorithm (DFA) for feature selection instead of continuous algorithms, which can be time-consuming when dealing with discrete face recognition.
1.2 Motivation.
Face recognition matches a person’s biometrics to a database of pre-defined faces to authenticate their identification. Face recognition involves feature extraction, selection, and categorization.
The feature selection process involves identifying the most representative feature subset for a given set. Researchers have proposed metaheuristic and hybrid search algorithms for feature selection.
Continuous algorithms require more recognition time and reduce accuracy. Discrete algorithms, consisting of extracted features, lead to efficient recognition time and accuracy.
Need help with a related project topic or New topic? Send Us Your Topic