Data Hiding (1 of 3)Review of Last ClassQuick Review on Image Compression, etc.What is An Image?Sampling and QuantizationExamples of SamplingExamples of QuantizationDifferent Color RepresentationsExamplesWhy Do Transforms?Review of 1-D & 2-D Unitary TransformsCommon Unitary TransformsLossless Coding ToolsTransform CodingIllustration of JPEG Baseline AlgorithmAdditive Data HidingCrypto is Useful, but Not Enough ……Multimedia Data Hiding / Digital WatermarkingGeneral FrameworkIssues and ChallengesAdditive Embedding: Basic IdeasTheoretical FoundationsSpread Spectrum Approach: Cox et al (NECI)Cox’s Scheme (cont’d)Slide 25Invisible Robust Wmk: Improved SchemesCompare Cox & Podilchuk SchemesCompare Cox & Podilchuk Schemes (cont’d)SummarySuggested readingQuestion for Today (QFT)Slide 32Type-I Additive EmbeddingType-II Relationship Enforcement EmbeddingConveying One-bit Through Noisy Channel (cont’d)Performance of Optimal DetectorRelated TerminologyCategories of WatermarkingMajor ApplicationsMajor Applications (cont’d)Watermarking vs. Data HidingVerify Ownership: Invisible Robust WmkECE738 Advanced Image ProcessingData Hiding (1 of 3)Curtsey of Professor Min Wu Electrical & Computer EngineeringUniv. of Maryland, College ParkMin Wu @ U. Maryland 20022ECE738 Advanced Image ProcessingReview of Last Class•Wrap up optimal 1-bit detection–Performance is determined by SNR and signal length (# observations)–Detection under low SNR ~ use longer signal•Cryptographic tools for secure communications–Building blocks: pseudo-random # generator, one-way func., hash–Encryption–Integrity verification (tampering detection)=> 3rd lecture notes http://www.ece.umd.edu/class/enee739m/lec/739S02_lec3.pdf•Today–Quick review on image processing–Intro. to data hiding: additive embeddingMin Wu @ U. Maryland 20023ECE738 Advanced Image ProcessingQuick Review on Image Compression, etc.Min Wu @ U. Maryland 20024ECE738 Advanced Image ProcessingWhat is An Image?•Grayscale image–A grayscale image is a function I(x,y) of the two spatial coordinates of the image plane.–I(x,y) is the intensity of the image at the point (x,y) on the image plane.–We can restrict the image to be bounded by some rectangle [0,a][0,b]I: [0, a]  [0, b]  [0, inf )•Color image–Can be represented by three functions, R(x,y) for red, G(x,y) for green, and B(x,y) for blue.Min Wu @ U. Maryland 20025ECE738 Advanced Image ProcessingSampling and Quantization•Computer handles “discrete” data.•Sampling–Sample the value of the image at the nodes of a regular grid on the image plane.–A pixel (picture element) at (i, j) is the image intensity value at grid point indexed by the integer coordinate (i, j).•Quantization–Is a process of transforming a real valued sampled image to one taking only a finite number of distinct values.–Each sampled value in a 256-level grayscale image is represented by 8 bits.0 (black)255 (white)Min Wu @ U. Maryland 20026ECE738 Advanced Image ProcessingExamples of Sampling256x25664x6416x16Min Wu @ U. Maryland 20027ECE738 Advanced Image ProcessingExamples of Quantization8 bits / pixel4 bits / pixel2 bits / pixelMin Wu @ U. Maryland 20028ECE738 Advanced Image ProcessingDifferent Color Representations•RGB•YIQ for NTSC transmission system–National Television Systems Committee (NTSC)–Receiver primary sys. (RN, GN, BN) as TV receivers standard–Transmission system (Y, I, Q)• facilitate transmission of color video via monochrome TV ch.•YUV (YCbCr) for PAL and digital video•HSV ~ Hue, Saturation, Value•CMY for printing–Cyan, Magenta, Yellow (complement of RGB)Min Wu @ U. Maryland 20029ECE738 Advanced Image ProcessingExamplesHSVYUVRGBMin Wu @ U. Maryland 200210ECE738 Advanced Image ProcessingWhy Do Transforms?•Fast computation–E.g., convolution vs. multiplication•Conceptual insights for various image processing–E.g., spatial frequency info. (smooth, moderate change, fast change, etc.)•Obtain transformed data as measurement–E.g., radiology images (medical and astrophysics)–Need inverse transform–May need to get assistance from other transforms•For efficient storage and transmission–Pick a few “representatives” (basis) –Just store/send the “contribution” from each basisMin Wu @ U. Maryland 200211ECE738 Advanced Image ProcessingReview of 1-D & 2-D Unitary Transforms•Vector/matrix representation of 1-D & 2-D sampled signal–Representing an image as a matrix or sometimes as a long vector•Basis functions/vectors and orthonormal basis–Used for representing the space via their linear combinations–Many possible sets of basis and orthonormal basis•Unitary transform on input x ~ A-1 = A*T –y = A x  x = A-1 y = A*T y =  ai*T y(i) ~ represented by basis vectors {ai*T}–Rows (and columns) of a unitary matrix form an orthonormal basis•General 2-D transform and separable unitary 2-D transform–2-D transform involves O(N4) computation–Separable: Y = A X AT = (A X) AT ~ O(N3) computation•Apply 1-D transform to all columns, then apply 1-D transform to rowsMin Wu @ U. Maryland 200212ECE738 Advanced Image ProcessingCommon Unitary Transforms–DFT, DCT, HaarSee also: Jain’s Fig.5.2 pp136Min Wu @ U. Maryland 200213ECE738 Advanced Image ProcessingLossless Coding Tools•PCM encoding–Fixed-length encoding of a sampled and quantized signal•Entropy encoding–Basic ideas ~ why bring in probability distribution?•Assign shorter codeword to commonly seen values–Limit of compression ~ Entropy –Huffman coding–Run-length coding•Predictive coding–Basic ideas and DPCMMin Wu @ U. Maryland 200214ECE738 Advanced Image ProcessingTransform Coding•Basic ideas–Energy compaction via appropriate transform–Adaptive bit allocation• allocate more bits to info.-rich coefficient bands•General block-based transform coding–Tradeoff for block size–Ordering & Zonal/Threshold coding•JPEG baseline algorithm (block DCT based)Min Wu @ U. Maryland 200215ECE738 Advanced Image ProcessingIllustration of JPEG Baseline Algorithm–Block diagram from Wallace’s JPEG tutorial paper–Flash demo by Dr. Ken Lam (Hong Kong PolyTech Univ.)Min Wu @ U. Maryland 200216ECE738 Advanced Image ProcessingAdditive Data HidingMin Wu @ U. Maryland 200217ECE738 Advanced Image ProcessingCrypto is Useful, but Not Enough ……•Encryption–Helps to protect confidentiality –Protection vanishes after decryption–Prefer a way to