An Adaptive Texture Enhancement Algorithm for AR Live Screen Based on Approximate Matching

In order to improve the visual effect of AR live video detail texture, an adaptive enhancement algorithm based on approximate matching is proposed. According local self-similarity original image, best block initial super-resolution image obtained. After extracting its high-frequency information, improved singular value decomposition (SVD) used embed watermark into gray image; And through extraction and reconstruction matrix, effective texture information obtained, then final completed. The experimental results show that peak signal-to-noise ratio, structure similarity feature are high, good; enhancement, clarity edge sharpness picture better, achieved.