What does this research mean for the field?

A proposed deep learning framework combining neural style transfer, adaptive color enhancement, and feature extraction effectively restores and enhances Chinese fine-art portraits while simultaneously preserving facial identity, structural consistency, texture realism, and artistic style. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

Develop an automated deep learning framework to enhance faded colors and degraded textures in Chinese fine-art portraits while preserving artistic style, structural consistency, and facial identity.

August 23, 2026International Journal of Software Engineering and Knowledge Engineering

Enhancing Artistic Portrait Colors Using Deep Learning and Style Transfer: An Information Literacy Approach

View Full Paper

Ask AI

Bookmark

Authors

XCXiong CaoHarbin Institute of Technology QKQiongqiong KangTwitter (United States)

Discussion

Member takes

Implication

Computational framework demonstrates high-fidelity color and texture restoration in historical Chinese fine-art portraits, suggesting robust utility for digital cultural heritage preservation.

Key Points

Develop an automated deep learning framework to enhance faded colors and degraded textures in Chinese fine-art portraits while preserving artistic style, structural consistency, and facial identity.
Trained and evaluated on the Chinese Fine Art/Famous Chinese Paintings Dataset using RetinaFace for face detection.
Combined Bilateral Filtering for edge-preserving denoising, Local Binary Patterns (LBP) for texture extraction, and VGG-19 for gradient-based Neural Style Transfer Optimization (GB-NSTO).
Implemented Adaptive Histogram Based Color Enhancement (AHCE) to adjust luminance and color saturation.
Achieved an SSIM of 0.92, a PSNR of 28.5 dB, an FID of 45.3, and a Colorfulness Index of 62.7.
Demonstrated reduced computational inference time compared to baseline methods while maintaining high artistic fidelity and facial identity preservation.

Cite This Study

Cao et al. (2026) studied this question.

synapsesocial.com/papers/6a8aada77677a34114445eb2 https://doi.org/10.1142/s0218194026500713

View Full Paper

Ask AI

Bookmark

Implication

Computational framework demonstrates high-fidelity color and texture restoration in historical Chinese fine-art portraits, suggesting robust utility for digital cultural heritage preservation.