Blind PSNR estimation of video sequences, through non-uniform quantization watermarking

This paper describes a watermark-based technique that estimates the PSNR of encoded video without requiring the original media data. Watermark embedding is performed in the block-based DCT domain using a new nonuniform quantization scheme that accounts for human vision characteristics. At the extraction, the square distances between received coefficients and nearest quantization points are computed, giving a frame-by-frame estimative for the mean square error (and consequently, the PSNR). Since these distances may be underestimated if distortion is greater than the quantization step used for watermark embedding, it is also proposed the use of the watermark extraction bit error rate, together with image statistics, for PSNR estimation weighting. Results show that PSNR estimation closely follows the true PSNR for a set of video sequences subject to different encoding rates.