Dimmed displays are commonly used to extend the battery life of electronic devices by reducing power consumption. Additionally, lower brightness levels help alleviate eye strain and discomfort during prolonged use. However, a key disadvantage of dimmed displays is the degradation in image quality due to reduced brightness and contrast. This paper presents an innovative approach that maintains the reduced power consumption of dimmed displays while enhancing the visual quality of the images shown. Our method modifies the image to exploit the human eye's sensitivity characteristics while keeping the display's overall brightness constant. By leveraging human contrast sensitivity and just-noticeable differences, the proposed algorithm improves image quality without increasing energy consumption. A novel technique for scaling just-noticeable difference thresholds using human frequency sensitivity filtering is introduced. Subjective tests revealed that participants preferred the modified images 70.54% of the time compared to the original dimmed images, demonstrating the effectiveness of our approach.

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

A Novel Image Enhancement Method for Dimmed Displays Based on JND and Human Frequency Sensitivity

  • Hamid Reza Tohidypour,
  • Panos Nasiopoulos,
  • Kristyn Lee,
  • Catherina Liu,
  • George Du,
  • Yubo Wang

摘要

Dimmed displays are commonly used to extend the battery life of electronic devices by reducing power consumption. Additionally, lower brightness levels help alleviate eye strain and discomfort during prolonged use. However, a key disadvantage of dimmed displays is the degradation in image quality due to reduced brightness and contrast. This paper presents an innovative approach that maintains the reduced power consumption of dimmed displays while enhancing the visual quality of the images shown. Our method modifies the image to exploit the human eye's sensitivity characteristics while keeping the display's overall brightness constant. By leveraging human contrast sensitivity and just-noticeable differences, the proposed algorithm improves image quality without increasing energy consumption. A novel technique for scaling just-noticeable difference thresholds using human frequency sensitivity filtering is introduced. Subjective tests revealed that participants preferred the modified images 70.54% of the time compared to the original dimmed images, demonstrating the effectiveness of our approach.