1、Digital Image Watermarking,Er-Hsien Fu EE381K-15280 Student Presentation,Overview,Introduction BackgroundWatermark PropertiesEmbeddingDetection The ProjectIntroductionEmbeddingDetection Conclusions,Introduction,Watermark-an invisible signature embedded inside an image to show authenticity or proof o
2、f ownership Discourage unauthorized copying and distribution of images over the internet Ensure a digital picture has not been altered Software can be used to search for a specific watermark,Background Watermark Properties,Watermark should appear random, noise-like sequence Appear Undetectable Good
3、Correlation PropertiesHigh correlation with signals similar to watermarkLow correlation with other watermarks or random noise Common sequencesA) Normal distributionB) m-sequences,W=1 0 0 1 00 1 1 0 11 1 0 1 00 1 1 1 10 1 0 0 0,Project: Introduction,Possible for watermark to be binary sequence Error-
4、correction coding techniques Use convolutional codes Decode by Viterbi algorithm Compare with non-coding method See if it improves watermark detection More or less robust to attacks?Additive noise, JPEG Compression, Rescale,Unzign Performance assessed by correlation coefficient,Watermark Embedding,W
5、atermark,Original Image,Watermarked image,Watermark placed into information content of Original Image to create Watermarked Image Image ContentSpatial Domain (Least Significant Bit)FFT - Magnitude and PhaseWavelet TransformsDCT Coefficients,Setup-Watermark Embedding,Image,1000 Highest Coeff,Conv Cod
6、e,DCT,Inter- leave,Water- mark,Water- marked Image,IDCT,DC Component Excluded for 1000 Highest Coefficients Interleaving prevents burst errors Watermarked Image Similar to original image Without coding, ignore Conv Code and Interleave block,Original Image,Watermarked Image, No Coding,Watermarked Ima
7、ge with Coding,512x512 “Mandrill” Image See Handout Both watermarks imperceptible Alterations to original imagedifficult to notice,Watermark Detection,*,= ,Suspected Image,Extracted Watermark,Original Watermark,Correlation,Watermark Extracted from Suspected Image Compute correlation of Extracted and
8、 Original Watermark Threshold correlation to determine watermark existence,Watermark Detection,Corrupted Image,Original Image,Extracted Watermark,Owners watermark,Correlation Coefficient,1000 Highest DCT Coeff,Deinterleave, Viterbi Decode,For no coding, deinterleave and decode block ignored =EW1*W2/
9、 EW12EW22 If W1=W2 then =1 if W1 and W2 are independent, then =0 if EW1=0 Corruptions are additive noise, JPEG Compression Image scaling, and UnZign,W2,W1,Convolutional Codes,Input=.1011010101100000000 G0 = 1 1 1 1 0 1 0 1 1 G1 = 1 0 1 1 1 0 0 0 1,C0,C1,Output C0 = conv(G0,Input); Output C1=conv(G1,
10、Input) Convolutional code implemented using linear shift registers Adds redundancy for error-correction Encoding/Decoding well researched Good coding performance, very popular,Viterbi Decoding,0,1,2,3,State, ,Find most likely path through trellis Begin and end at all zero state Upper arrows = input=
11、0, Lower arrow =input=1 Every possible input/output combination is compared with the received output Optimal Decoding Method,With Coding: Additive Noise (0,900),No Coding: Additive Noise(0,900),Zero mean additive noise, variance=100, 400, 900 Both methods had high correlation Coding method performed
12、 slightly better For variance = 900 (no coding) = 77% p (coding) = 84%,4:1 JPEG Compression, No coding,4:1 JPEG Compression With Coding,JPEG Compression: 1.4:1, 2.2:1, 4:1 ratio Both methods resistant to JPEG compression Coding method outperformed non-coding method Perfect detection for coding metho
13、d,Watermark removal using Unzign,Convert to grayscale and resize,Unzign-watermark removal software Image resized to 512x512 and convert to grayscale before detection Moderate detection for without coding: (no coding) = 57% (coding) = 23% Coding method sensitive to resizing,Conclusions,Convolutional coding more immune to additive noise and JPEG Compression Coding method fragile w.r.t. rescaled images Moderate detection levels for unzigned images Further Suggestion: Try block DCT Use Wavelet Transform Exploit Human Visual System,Questions,
