Errata from "Fractal Image Compression: Theory and Application"

1. On page 9, Formula 1.2 should have the intergrand squared. (Thanks to Felix Kestenholz-Seiler for noticing that.)
2. On page 43, the formula for the rms metric is wrong. It is missing the "mean" part. The references to d_rms on that page should be changed to simply d. The rms metric is d_rms(. , .) = d(. , .)/sqrt(n). (Thanks to John T. MacCarthy for pointing this out).
3. Page 26, last line - the 2^2/A/epsilon requires a square on the epsilon. Similarly, on page 28, top line, the epsilon should be cubed. Further, Figure 2.1(d) could be drawn with three initial squares (rather than four) to better conform to definition 2.2, which states that we want to cover sets with the smallest number of boxes. (The dimension is the same if the boxes are on a lattice as is shown). (Thanks to John T. MacCarthy for pointing this out).
4. Page 74. The images in 3.19a and 3.19b should be exchanged.
5. Daniele Sparano (sparano@indigo.iet.unipi.it) writes: There is an error in equation 8.26 on the top of page 160. In the denominator the last term should also be raised to the power of 2.
6. Footnote 8 on page 45 contains an error. The color information is not sent during the color burst that occurs during the horizontal sync signal. Rather, the color data is amplitude modulated at around three and a half MHz, in two phases, and added to the Y signal. The color burst is only a synchronization signal, needed when color TVs first come out in order to synchronize the oscillators used to extract the color data. (Thanks to Scott Auchmoody, of ADI, Inc. for pointing this out).
7. On Pages 63 and 64, Figures 3.5 and 3.6 state that the image used was a 256x256. In fact, it was a 512x512 image. (My thanks to Lars Lundheim for catching this).
8. Page 294, Exercise 8 is missing a square root and a square inside the square root. It should read
   ||A|| = 1/2 ( .... -4(ad-bc)^2 ) ^ 1/2.
   My thanks to Sandeep Sikka for pointing some of this out.
9. On page 148, line 9: the 1/r should be sqrt(1/r).
   My thanks to Wei Xu from U. of Texas for pointing this out.
10. There is an error in eqs. of \beta1 and \beta2 on page 150. In the denominators, a minus before the squared \sums is missing. (Thanks to Franck Davoine for pointing this out):
11. Thanks to Hadi A Murr for the following:
    • Page 6, 4th line from the bottom: "representinging" .
    • Page 16, 9th line from the bottom: "Lets" should be "Let's" .
    • Page 80, Paragrph 4, line 8, "of of" .
    • Page 80, 7th line from the bottom: "verstorsprecisely" .
    • Page 126, top. The formula should read "a' = ..."
    • Page 128, last line: "with a cost of with a cost of" .