Multimedia Technology Lecture 7: Lossy Compression Algorithms

Multimedia Compression

• The most common type of multimedia compression algorithm is lossy compression.

Characteristics of Lossy Compression

• The main characteristic of lossy compression is that it discards some of the data, resulting in a loss of quality.

Image Compression

• JPEG (Joint Photographic Experts Group) is commonly used for digital images.
• JPEG typically achieves a compression ratio of 10:1 to 20:1 for digital images.

Distortion and Compression

• Distortion refers to the difference between the original and reconstructed data after compression.

Selectable Compression

• Transform coding allows for a selectable tradeoff between storage size and image quality.

Mean Square Error (MSE)

• The formula for MSE in image compression is: MSE = (1/N) * Σ(xi - yi)^2, where xi is the original data and yi is the reconstructed data.

Quality Assessment

• Peak Signal-to-Noise Ratio (PSNR) is the most commonly used measure to assess the quality of reconstruction of lossy compression codecs in image compression.

PSNR Values

• Typical PSNR values for lossy image and video compression with 8-bit data are 30-40 dB.
• Typical PSNR values for lossy image and video compression with 16-bit data are 40-50 dB.

Frequency Domain

• The frequency domain is where we deal with the rate at which pixel values are changing.

Discrete Cosine Transform (DCT)

• DCT is considered a better representation for image data in the frequency domain because it is more efficient and effective in representing the spatial frequencies of an image.
• The rationale behind transform coding using DCT is that it takes advantage of the fact that most of the image's energy is concentrated in the lower frequencies.

PSNR and MSE

• PSNR stands for Peak Signal-to-Noise Ratio in image compression.
• MSE stands for Mean Square Error in image compression.
• The average square value of the original data sequence in SNR (Signal-to-Noise Ratio) calculation is the signal power.