DEA-Net: Single Image Dehazing Based on Detail-Enhanced Convolution and Content-Guided Attention

Single image dehazing is a challenging ill-posed problem which estimates latent haze-free images from observed hazy images. Extensive experimental results demonstrate the effectiveness of our DEA-Net, outperforming the advanced (SOTA) methods by boosting the PSNR index over 41 dB with only 3.653 M parameters.

⚡ This is an original paraphrased summary — not copied from the abstract. Full paper available at the source link below.

• 1 Some existing deep learning based methods are devoted to improving the model performance via increasing the depth or width of convolution.
• 2 The learning ability of Convolutional Neural Network (CNN) structure is still under-explored.
• 3 In this paper, a Detail-Enhanced Attention Block (DEAB) consisting of Detail-Enhanced Convolution (DEConv) and Content-Guided Attention (CGA) is proposed to boost the feature learning for improving the dehazing performance.

This research advances how AI systems learn, reason, and solve problems — with direct implications for automation and scientific discovery.

This summary is based on publicly available metadata and abstract. For the full research paper, visit the original source: