ic-proj2

Second Lab Work of IC Course

This repository contains two independent components:

1. Golomb Coding — A C++ implementation of Golomb encoding/decoding for integers.
2. Extract Color Channel — An OpenCV-based image utility to extract a single color channel.

Both can be built from the same Makefile using different targets.

---

🧱 Build

On the root project directory, you can build everything or each target separately:

Build both (default)

make

Build only the OpenCV example

make extract

Build only the Golomb coder

make golomb

Build only the Image Codec

make image_codec

The binaries are placed inside the build/ directory:

• build/golomb
• build/extract_color_channel
• build/image_transform
• build/image_codec

---

▶️ Run

Exercise 1 — Extract Color Channel (OpenCV)

./build/extract_color_channel <input_image> <output_image> <channel_index>

Channel index (BGR):

• 0 → Blue
• 1 → Green
• 2 → Red

Example:

./build/extract_color_channel images-ppm/airplane.ppm airplane_red.pgm 2

Exercise 2 — Image Transformations (pixel-by-pixel)

Create negatives, mirrors, rotate by multiples of 90°, and brightness changes.

Usage:

./build/image_transform <input> <output> <operation> [param]

Operations:

• neg — negative
• mirror_h — horizontal mirror
• mirror_v — vertical mirror
• rotate <k> — rotate by k*90 degrees (k integer)
• bright <delta> — add delta to all channels (delta can be negative)

Examples:

./build/image_transform images-ppm/lena.ppm build/lena_neg.ppm neg
./build/image_transform images-ppm/lena.ppm build/lena_rot90.ppm rotate 1
./build/image_transform images-ppm/lena.ppm build/lena_bright_plus50.ppm bright 50

---

Exercise 3 — Golomb Coding

You can encode or decode integer sequences using Golomb coding. Run with:

./build/golomb -m <m_value> -mode <mode> encode <list_of_integers>

Parameters:

• -m <m_value> → the Golomb parameter (must be ≥ 1)

• -mode <mode> → negative number handling mode:

  • sign → sign-and-magnitude
  • interleave → positive/negative interleaving (zig-zag)

Example (interleaved mode):

./build/golomb -m 3 -mode interleave encode 0 -1 5 10

---

Exercise 4 - Golomb Codec

You can losslessly compress audio files into custom .glb format. Usage:

./golomb_codec encode input.wav output.glb

or:

./golomb_codec decode output.glb output.wav

---

Exercise 5 — Image Codec

Lossless grayscale image codec based on spatial prediction and Golomb coding of prediction residuals. The tool expects a single-channel 8-bit image as input (PGM or grayscale PNG). If you start from a color image, extract one channel first (see Exercise 1).

Usage:

# Encode (mode must come first). Predictor: 0=left, 1=median (default=1)
./build/image_codec encode <input_gray> <output.gimg> [predictor]

# Decode to any image format supported by OpenCV (e.g., PNG)
./build/image_codec decode <input.gimg> <output_image>

Notes:

• Input must be single-channel 8-bit; convert or extract a channel if needed.
• If you omit encode/decode as the first argument, you'll get "Unknown mode".
• The encoder tries several m values and prints the chosen parameter and bit count.
• Decoding is lossless (pixel-by-pixel identical to the original input).

---

🧹 Clean

To remove the compiled binaries:

make clean

---

⚙️ Notes

• The Makefile automatically checks for OpenCV via pkg-config. If the OpenCV package isn’t found, you can specify one manually:

  make extract PKG=opencv

• The Golomb module has no external dependencies and can always be built.

---