Mumbai University > EXTC > Sem 7 > Data Compression and Encryption

Marks: 5 M

Year: Dec 2014

i. If the data we are attempting to compress consists of numerical values, such as images, using context-based approaches directly can be problematic.

ii. There are several reasons for this. Most context-based schemes exploit exact reoccurrence of patterns.

iii. Images are usually acquired using sensors that have a small amount of noise. While this noise may not be perceptible, it is sufficient to reduce the occurrence of exact repetitions of patterns.

iv. A simple alternative to using the context approach is to generate a prediction for the value to be encoded and encode the prediction error.

v. If there is considerable dependence among the values with high probability, the prediction will be close to the actual value and the prediction error will be a small number.

vi. We can encode this small number, which as it occurs with high probability will require fewer bits to encode

vii. By plotting histogram of the difference between neighbouring pixels. we can observe that small differences occur much more frequently and therefore can be encoded using fewer bits.

viii. Because of the strong correlation between pixels in a neighbourhood, predictive coding has been highly effective for image compression. It is used in the JPEG-LS, which is the standard for lossless image compression.

ix. The function of the predictor is to obtain an estimate of the current sample based on the reconstructed values of the past sample.

x. The requirement that the prediction algorithm use only the reconstructed values is to ensure that the prediction at both the encoder and the decoder are identical.

xi. The reconstructed values used by the predictor, and the prediction algorithm, are dependent on the nature of the data being encoded.

For example, for speech coding the predictor often uses the immediate past several values of the sequence, along with a sample that is a pitch period away, to form the prediction.