A Context-Based Image Contrast Enhancement Using Energy Equalization With Clipping Limit

Area of Science:

• Computer Vision
• Image Processing
• Artificial Intelligence

Background:

• Traditional histogram equalization (HE) can lead to over-enhancement and noise amplification.
• Existing contrast enhancement methods often fail to fully utilize spatial information within images.

Purpose of the Study:

• To propose a new context-based image contrast enhancement technique called energy curve equalization (ECE).
• To improve image quality by addressing limitations of conventional histogram equalization.

Main Methods:

• Developed a novel contrast enhancement process using energy curve equalization (ECE) with a clipping limit.
• Utilized a modified Hopfield neural network architecture to compute the energy curve, incorporating spatial adjacency information.
• Subdivided the clipped energy curve into three regions for individual equalization, creating transfer functions for the final enhanced image.

Main Results:

• The proposed ECE method produces a smoother energy curve compared to traditional histograms.
• A clipping limit, set as the average of mean and median, effectively prevents over-enhancement.
• Qualitative and quantitative assessments demonstrate the efficiency of ECE compared to conventional histogram equalization techniques.

Conclusions:

• ECE offers a superior approach to image contrast enhancement by leveraging spatial information and a smoother energy curve.
• The clipping limit and regional equalization contribute to more effective and controlled image enhancement.
• The proposed method shows significant potential for improving image quality in various applications.