---

TOPIC IMPLEMENT A BASIC IMAGE PROCESSING PIPELINE

1. INITIALS load, check, and convert into a double-precision array

bits per integer? 16 bits - uint16

width? 4290

height? 2856

Print them using fprintf function.

2. LINEARIZATION shift and scale, within the range [0, 1]

Linear transform mapped.

3. IDENTIFYING THE CORRECT BAYER PATTERN identify Bayer patterns

version? rggb

reason?
I checked all results image of four possible red-green-blue patterns.
Result of rggb is best representor of original image.
And also, I calculated sum of differences between pixels, and I figured out difference betweeen img_2 and img_3 is smallest.
It means img_2 and img_3 represents same color, so Bayer pattern of this camera is rggb.

Calculate img1, img2, img3, img4. And generate test images with cat function

4. WHITE BALANCING white world and gray world automatic white balancing

best? white world automatic white balancing

Implemented two functions: white_balancing_grey and white_balancing_white.

5. DEMOSAICING use bilinear interpolation

Just use interp2 function.

6. BRIGHTNESS ADJUSTMENT AND GAMMA CORRECTION apply non-linear gamma correction

Implement given non-linear function to apply gamma correction.

7. COMPRESSION PNG vs JPEG, and compression ratio

difference between .PNG and .JPEG(quality 95)? honestly, nope

size of the uncompressed file?(.PNG) 14,439,587 bytes

size of the compressed file?(.JPEG, 95) 2,679,370 bytes

compression ratio? 2,679,370/14,439,587 = 0.1856

compression ratio of quality 50? 608,555/14,439,587 = 0.0421

compression ratio of quality 30? 452,203/14,439,587 = 0.0313

compression ratio of quality 20? 366,003/14,439,587 = 0.0253

compression ratio of quality 15? 321,151/14,439,587 = 0.0222

compression ratio of quality 10? 273,577/14,439,587 = 0.0189

compression ratio of quality 5? 222,342/14,439,587 = 0.0154

the lowest setting for indistinguishable, compression ratio? quality 20 -> compression ratio 0.0253

Used imwrite function.

You can check full matlab codes in GitHub Repo.

---