DIGITAL IMAGE PROCESSING Part I: Image Transforms 1-D SIGNAL TRANSFORM GENERAL FORM • Scalar form { f ( x), 0 x N 1} N 1 g (u) T (u, x) f ( x) x 0 • Matrix form g T f 0 u N 1 1-D SIGNAL TRANSFORM cont. REMEMBER THE 1-D DFT!!! • General form { f ( x), 0 x N 1} N 1 g (u) T (u, x) f ( x) x 0 0 u N 1 • DFT N 1 1 g (u ) x 0 N ux j 2 N e f ( x) 1-D INVERSE SIGNAL TRANSFORM GENERAL FORM • Scalar form N 1 f ( x) I ( x, u ) g (u ) u 0 • Matrix form 1 f T g 1-D INVERSE SIGNAL TRANSFORM cont. REMEMBER THE 1-D DFT!!! • General form { f ( x), 0 x N 1} N 1 f ( x) I ( x, u ) g (u ) u 0 • DFT ux N 1 j 2 N g (u ) e f ( x) x 0 0 u N 1 1-D UNITARY TRANSFORM • Matrix form g T f T 1 T T SIGNAL RECONSTRUCTION T 1 T T T N 1 f T g f ( x) T (u, x) g (u) u 0 IMAGE TRANSFORMS • Many times, image processing tasks are best performed in a domain other than the spatial domain. • Key steps: (1) Transform the image (2) Carry the task(s) in the transformed domain. (3) Apply inverse transform to return to the spatial domain. 2-D (IMAGE) TRANSFORM GENERAL FORM N 1 N 1 g (u , v) T (u , v, x, y ) f ( x, y ) x 0 y 0 N 1N 1 f ( x, y ) I ( x, y, u, v) g (u, v) u 0 v 0 2-D IMAGE TRANSFORM SPECIFIC FORMS • Separable T (u, v, x, y) T1 (u, x)T2 (v, y) • Symmetric T (u, v, x, y) T1 (u, x)T1 (v, y) • Separable and Symmetric g T1 f T T 1 • Separable, Symmetric and Unitary f T T1 g T1 ENERGY PRESERVATION • 1-D g f 2 2 • 2-D M 1 N 1 2 M 1 N 1 f ( x, y ) g (u , v) x 0 y 0 u 0 v 0 2 ENERGY COMPACTION Most of the energy of the original data concentrated in only a few of the significant transform coefficients; remaining coefficients are near zero. Why is Fourier Transform Useful? • Easier to remove undesirable frequencies. • Faster to perform certain operations in the frequency domain than in the spatial domain. • The transform is independent of the signal. Example Removing undesirable frequencies noisy signal frequencies To remove certain frequencie s, set their correspond ing F (u ) coeeficien ts to zero! remove high frequencies reconstructed signal How do frequencies show up in an image? • Low frequencies correspond to slowly varying information (e.g., continuous surface). • High frequencies correspond to quickly varying information (e.g., edges) Original Image Low-passed 2-D DISCRETE FOURIER TRANSFORM M 1N 1 F (u, v) f ( x, y)e j 2 (ux / M vy / N ) x 0 y 0 1 M 1N 1 j 2 (ux / M vy / N ) f ( x, y) F (u, v)e MN u 0 v 0 Visualizing DFT • Typically, we visualize F (u, v) • The dynamic range of F (u, v) is typically very large • Apply stretching: D(u, v) c log[1 F (u, v) ] ( c is constant) original image before scaling after scaling Amplitude and Log of the Amplitude Amplitude and Log of the Amplitude Original and Amplitude DFT PROPERTIES: SEPARABILITY Rewrite F (u, v) as follows: M 1 F (u , v) e j 2ux / M x 0 N 1 f ( x, y )e y 0 If we set: N 1 G ( x, v ) f ( x, y )e j 2vy / N y 0 Then: M 1 F (u , v) e x 0 j 2ux / M G ( x, v ) j 2vy / N DFT PROPERTIES: SEPARABILITY • How can we compute G( x, v) ? N 1 G ( x, v ) f ( x, y ) e j 2vy / N y 0 1 N 1 j 2vy / N N f ( x, y ) e N y 0 N DFT of rows of f ( x. y ) • How can we compute F (u, v) ? 1 M 1 j 2ux / M F (u, v) M G( x, v)e M x 0 M DFT of columnsof G( x, v) DFT PROPERTIES: SEPARABILITY DFT PROPERTIES: PERIODICITY The DFT and its inverse are periodic with period N F (u, v) F (u M , v) F (u, v N ) F (u M , v N ) DFT PROPERTIES: SYMMETRY If F (u, v) is real, then F (u , v) F (u ,v) F (u, v) F (u ,v) * f ( x, y ) real and even F (u, v) real and even f ( x, y ) real and odd F (u , v) imaginary and odd DFT PROPERTIES: TRANSLATION • Translation in spatial domain: f ( x x0 , y y0 ) F (u, v)e j 2 (u0 x / M v0 y / N ) • Translation in frequency domain: f ( x, y)e j 2 (u0 x / M v0 y / N ) F (u u0 , v v0 ) DFT PROPERTIES: TRANSLATION Warning: to show a full period, we need to translate the origin of the transform at u (u, v) ( M / 2, N / 2) N /2 F (u ) N F (u ) 2 DFT PROPERTIES: TRANSLATION To move F (u, v) at ( M / 2, N / 2) take u 0 M / 2 and v0 N / 2 In that case j 2 (u0 x / M v0 y / N ) e Using f ( x, y )e e j ( x y ) j 2 (u0 x / M v0 y / N ) f ( x, y )(1) ( x y) (1) ( x y) F (u u0 , v v0 ) F (u M / 2, v n / 2) DFT PROPERTIES: TRANSLATION f ( x, y)(1) ( x y) F (u M / 2, v n / 2) no translation after translation DFT PROPERTIES: ROTATION rotating f ( x, y) by rotates F (u, , v) by DFT PROPERTIES ADDITION-MULTIPLICATION [ f ( x, y) g ( x, y)] [ f ( x, y)] [ g ( x, y)] [ f ( x, y) g ( x, y)] [ f ( x, y)] [ g ( x, y)] where [] is the Fourier transform DFT PROPERTIES: SCALE [ f ( x, y) g ( x, y)] [ f ( x, y)] [ g ( x, y)] [ f ( x, y) g ( x, y)] [ f ( x, y)] [ g ( x, y)] where [] is the Fourier transform DFT PROPERTIES: AVERAGE Average value of the image : 1 M 1 N 1 f ( x, y ) f ( x, y ) MN x 0 y 0 M 1 N 1 F (0,0) f ( x, y ) x 0 y 0 1 F (0,0) f ( x, y ) MN Original Image Fourier Amplitude Fourier Phase Magnitude and Phase of DFT • What is more important? magnitude phase • Hint: use inverse DFT to reconstruct the image using magnitude or phase only information Magnitude and Phase of DFT Reconstructed image using magnitude only (i.e., magnitude determines the contribution of each component!) Reconstructed image using phase only (i.e., phase determines which components are present!) Magnitude and Phase of DFT Original Image-Fourier Amplitude Keep Part of the Amplitude Around the Origin and Reconstruct Original Image (LOW PASS filtering) Keep Part of the Amplitude Far from the Origin and Reconstruct Original Image (HIGH PASS filtering) Reconstruction from phase of one image and amplitude of the other Reconstruction from phase of one image and amplitude of the other