Adaptive total variation minimization-based image enhancement from flash and no-flash pairs

Area of Science:

• Computer Vision
• Image Processing
• Computational Photography

Background:

• Image quality is often degraded by noise or lack of detail.
• Capturing images with both flash and no-flash provides complementary information.
• Combining information from multiple images is a key challenge in image enhancement.

Purpose of the Study:

• To develop a novel image enhancement technique.
• To improve image quality by merging flash and no-flash image data.
• To achieve efficient image denoising while preserving important image details.

Main Methods:

• A novel approach combining flash and no-flash images.
• Utilizing Adaptive Total Variation Minimization (ATVM).
• Preserving strong gradients in the flash image for detail retention.

Main Results:

• The proposed method effectively enhances image quality.
• Demonstrated efficient image denoising capabilities.
• Successfully preserved strong image gradients.

Conclusions:

• The novel approach offers a significant improvement in image enhancement.
• Adaptive Total Variation Minimization proves effective for denoising.
• The technique successfully merges ambient light and flash details for superior image quality.