Understanding JPEG Compression: Lossy Vs. Lossless Techniques.pdf

ON COMPRESSION by Cory Arcangel (2007-2008)

This article explains how JPEG image compression utilizes the Discrete Cosine Transform (DCT).

Key Concepts:

Lossy vs. Lossless Compression:
Lossless: Preserves all original data, like zip files.
Lossy: Discards some data to achieve smaller file sizes, commonly used for images, music, and video.

Discrete Cosine Transform (DCT): A mathematical formula that breaks down an image into frequency components, highlighting which parts are most important.

* Quantization: The process of limiting the number of possible values for each DCT coefficient, effectively discarding less significant information.

Process:

1. DCT Formula: Converts an image (matrix) into a set of DCT coefficients that represent the contribution of different frequencies to the image.

2. Comparison & Coefficient Generation: The input image is compared to basis functions (cosine curves of increasing frequency) to generate DCT coefficients reflecting the presence of each frequency in the image.

3. Quantization: Discards less significant DCT coefficients, reducing the number of bits needed to store them.

4. Decompression: Reverses the process, using the remaining quantized coefficients to reconstruct the original image.

JPEG's "Blocky" Look:

The DCT, when combined with quantization, emphasizes certain frequencies in an image, often resulting in visible blocks of pixelated data, a characteristic "JPEG look." This is a trade-off for achieving smaller file sizes.

Standardization:

JPEG became a widely adopted standard for image compression due to the work of the Joint Photographic Experts Group (JPEG).