2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction
| Format: | General Document |
|---|
| _version_ | 1860798151927005184 |
|---|---|
| building | INTELEK Repository |
| collection | Online Access |
| collectionurl | https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 |
| copyright | Copyright©PWB2025 |
| country | Malaysia |
| date | 2022-09-04 |
| format | General Document |
| id | 16198 |
| institution | UniSZA |
| originalfilename | 16198_50e4f9ef74755a4.pdf |
| person | Alsarayreh Ayah Mohammad Hussein |
| recordtype | oai_dc |
| resourceurl | https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16198 |
| sourcemedia | Server storage Scanned document |
| spelling | 16198 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16198 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu Faculty of Informatics & Computing English application/pdf 1.5 Server storage Scanned document Universiti Sultan Zainal Abidin UniSZA Private Access UNIVERSITI SULTAN ZAINAL ABIDIN SAMBox 2.3.4; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) 279 Copyright©PWB2025 2022-09-04 Face—Recognition 16198_50e4f9ef74755a4.pdf Constrained Local Model (CLM) Alsarayreh Ayah Mohammad Hussein 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction Gender prediction based on face images has numerous potential contributions especially in image retrieval. Gender Predection systems can assist in tracking moving items, detecting anomalous activities, and facilitating the security investigation of criminals who purposefully try to disguise their identity information. There are numerous techniques exist for gender prediction, however their performance is lower than other pattern recognition tasks such as face recognition, age estimation, and iris recognition. Numerous approaches in the associated literature are problematic for gender estimation due to generalization issues such as accuracy and performance. Constrained Local Models (CLMs) are a well-established family of methods for facial landmark detection. However, CLM local detectors are unable to simulate the complicated individual landmark appearance impacted by emotion, illumination, facial hair, and makeup. Therefore, this study aims to improve the CLM in order to simulate the complicated landmarks. An integrated facial feature extraction model was proposed to improve the performance of CLM in obtaining the best result for feature detection. Three-patch expert response maps were used which are the face contour, nose ridge, and chin. The fitting of a response image to a 2-D quadratic was solved via quadratic function with constraints. Two publicly accessible databases namely LFBW and MORPH were tested in experiments. Each dataset consisted of 490 training images and 210 test images. The machine learning algorithm was then used for gender prediction. A hierarchical approach was developed to train the machine learning algorithms for the CLM of facial features with gender labels. Next, the proposed framework was developed to jointly address the issues of categorizing the CLM test features into gender labels based on the trained data. The performance of the proposed model was evaluated based on the Mean Absolute Error (MAE), Cumulative Score (CS), and processing time. Experimental results indicated that the proposed model is effective in extracting facial features and predicting gender. The proposed model provides the best MAE value of 3.64% for the LFPW and MORPH databases through Progressive Transudative Support Vector Machines (PTSVM). This result is more promising than the results of several state-of-the-art techniques, such as CLM using Constrained Local Neural Fields (CLNF) and CLM using Support Vector Regression (SVR), which provided the best MAE values of 4.24% and 4.74% for the LFPW respectively, and 4.64% and 4.11% for the MORPH respectively. The proposed model showed significant improvement in accuracy compared to previous studies. Three patch experts and a quadratic function for fitting the response image implies enhancing feature extraction. Consequently, the improvement done by this study gives significant results for gender prediction. Dissertations, Academic Facial Feature Extraction Gender Prediction Thesis |
| spellingShingle | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| state | Terengganu |
| subject | Face—Recognition Dissertations, Academic |
| summary | Gender prediction based on face images has numerous potential contributions especially in image retrieval. Gender Predection systems can assist in tracking moving items, detecting anomalous activities, and facilitating the security investigation of criminals who purposefully try to disguise their identity information. There are numerous techniques exist for gender prediction, however their performance is lower than other pattern recognition tasks such as face recognition, age estimation, and iris recognition. Numerous approaches in the associated literature are problematic for gender estimation due to generalization issues such as accuracy and performance. Constrained Local Models (CLMs) are a well-established family of methods for facial landmark detection. However, CLM local detectors are unable to simulate the complicated individual landmark appearance impacted by emotion, illumination, facial hair, and makeup. Therefore, this study aims to improve the CLM in order to simulate the complicated landmarks. An integrated facial feature extraction model was proposed to improve the performance of CLM in obtaining the best result for feature detection. Three-patch expert response maps were used which are the face contour, nose ridge, and chin. The fitting of a response image to a 2-D quadratic was solved via quadratic function with constraints. Two publicly accessible databases namely LFBW and MORPH were tested in experiments. Each dataset consisted of 490 training images and 210 test images. The machine learning algorithm was then used for gender prediction. A hierarchical approach was developed to train the machine learning algorithms for the CLM of facial features with gender labels. Next, the proposed framework was developed to jointly address the issues of categorizing the CLM test features into gender labels based on the trained data. The performance of the proposed model was evaluated based on the Mean Absolute Error (MAE), Cumulative Score (CS), and processing time. Experimental results indicated that the proposed model is effective in extracting facial features and predicting gender. The proposed model provides the best MAE value of 3.64% for the LFPW and MORPH databases through Progressive Transudative Support Vector Machines (PTSVM). This result is more promising than the results of several state-of-the-art techniques, such as CLM using Constrained Local Neural Fields (CLNF) and CLM using Support Vector Regression (SVR), which provided the best MAE values of 4.24% and 4.74% for the LFPW respectively, and 4.64% and 4.11% for the MORPH respectively. The proposed model showed significant improvement in accuracy compared to previous studies. Three patch experts and a quadratic function for fitting the response image implies enhancing feature extraction. Consequently, the improvement done by this study gives significant results for gender prediction. |
| title | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| title_full | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| title_fullStr | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| title_full_unstemmed | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| title_short | 2022_Improvement Of Constrained Local Model In Extracting Facial Features For Gender Prediction |
| title_sort | 2022_improvement of constrained local model in extracting facial features for gender prediction |