Scholarlink

<p>This paper introduces a novel face detection technique designed to address the challenges of occlusion and uneven illumination in still images. Accurate face detection is critical for applications such as facial recognition, gesture analysis, and surveillance systems, which require machines to perform tasks with human-like precision. The proposed method integrates multiple colour models—YCbCr, HSV, and L × a × b—to enhance face detection accuracy under difficult conditions, including partial obstructions and inconsistent lighting. To validate the approach, experiments were conducted using images from established public datasets, including the AR face dataset and the Colour FERET dataset. By employing advanced machine learning techniques, the method emulates human perceptual abilities, effectively detecting, localizing, and recognizing faces across various scenarios. The results highlight the method's robustness in overcoming limitations faced by traditional detection techniques. This research advances the capability of automated systems in face detection, offering potential improvements in biometric security, human–computer interaction, and intelligent video surveillance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Face Detection Technique for Challenging Conditions Using Multi-color Model Integration

  • 期刊论文

  • 发表日期：

    2025年9月9日

查看全文
  • Anupam Yadav,
  • Gadug Sudhamsu,
  • Navdeep Kaur,
  • Jayant Jagtap,
  • Mandeep Kaur Chohan,
  • MJanaki Ramudu,
  • Ish Kapila,
  • Ahmed Alkhayyat

摘要

This paper introduces a novel face detection technique designed to address the challenges of occlusion and uneven illumination in still images. Accurate face detection is critical for applications such as facial recognition, gesture analysis, and surveillance systems, which require machines to perform tasks with human-like precision. The proposed method integrates multiple colour models—YCbCr, HSV, and L × a × b—to enhance face detection accuracy under difficult conditions, including partial obstructions and inconsistent lighting. To validate the approach, experiments were conducted using images from established public datasets, including the AR face dataset and the Colour FERET dataset. By employing advanced machine learning techniques, the method emulates human perceptual abilities, effectively detecting, localizing, and recognizing faces across various scenarios. The results highlight the method's robustness in overcoming limitations faced by traditional detection techniques. This research advances the capability of automated systems in face detection, offering potential improvements in biometric security, human–computer interaction, and intelligent video surveillance.