Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Frequent Pattern Mining approach to Image Compression

Avinash Kadimisetty, C. Oswald, B. Sivalselvan

TL;DR

This work addresses the demand for efficient lossy image compression by introducing a CFSM-based approach that replaces the JPEG DCT stage with a combination of $k$-means clustering and closed frequent sequence mining, followed by Huffman encoding. By deriving a compact code table from closed frequent sequences and employing a modified support to avoid overlapping counts, the method achieves substantial compression gains while preserving visual quality, as evidenced on standard 512×512 images. The key contributions are the integration of clustering with CFSM (and $s_{mod}$) for encoding, and the demonstration of about a 45% improvement in compression ratios with negligible degradation in PSNR and SSIM. This approach offers a promising direction for pattern-based image compression with practical impact on storage and bandwidth efficiency.

Abstract

The paper focuses on Image Compression, explaining efficient approaches based on Frequent Pattern Mining(FPM). The proposed compression mechanism is based on clustering similar pixels in the image and thus using cluster identifiers in image compression. Redundant data in the image is effectively handled by replacing the DCT phase of conventional JPEG through a mixture of k-means Clustering and Closed Frequent Sequence Mining. To optimize the cardinality of pattern(s) in encoding, efficient pruning techniques have been used through the refinement of Conventional Generalized Sequential Pattern Mining(GSP) algorithm. We have proposed a mechanism for finding the frequency of a sequence which will yield significant reduction in the code table size. The algorithm is tested by compressing benchmark datasets yielding an improvement of 45% in compression ratios, often outperforming the existing alternatives. PSNR and SSIM, which are the image quality metrics, have been tested which show a negligible loss in visual quality.

Frequent Pattern Mining approach to Image Compression

TL;DR

This work addresses the demand for efficient lossy image compression by introducing a CFSM-based approach that replaces the JPEG DCT stage with a combination of -means clustering and closed frequent sequence mining, followed by Huffman encoding. By deriving a compact code table from closed frequent sequences and employing a modified support to avoid overlapping counts, the method achieves substantial compression gains while preserving visual quality, as evidenced on standard 512×512 images. The key contributions are the integration of clustering with CFSM (and ) for encoding, and the demonstration of about a 45% improvement in compression ratios with negligible degradation in PSNR and SSIM. This approach offers a promising direction for pattern-based image compression with practical impact on storage and bandwidth efficiency.

Abstract

The paper focuses on Image Compression, explaining efficient approaches based on Frequent Pattern Mining(FPM). The proposed compression mechanism is based on clustering similar pixels in the image and thus using cluster identifiers in image compression. Redundant data in the image is effectively handled by replacing the DCT phase of conventional JPEG through a mixture of k-means Clustering and Closed Frequent Sequence Mining. To optimize the cardinality of pattern(s) in encoding, efficient pruning techniques have been used through the refinement of Conventional Generalized Sequential Pattern Mining(GSP) algorithm. We have proposed a mechanism for finding the frequency of a sequence which will yield significant reduction in the code table size. The algorithm is tested by compressing benchmark datasets yielding an improvement of 45% in compression ratios, often outperforming the existing alternatives. PSNR and SSIM, which are the image quality metrics, have been tested which show a negligible loss in visual quality.
Paper Structure (12 sections, 5 equations, 3 figures, 4 tables, 4 algorithms)

This paper contains 12 sections, 5 equations, 3 figures, 4 tables, 4 algorithms.

Table of Contents

  1. Abstract
  2. Introduction
  3. Related Work
  4. Proposed Image Compression Algorithm
  5. Encoder
  6. Image Acquisition
  7. Image Clustering
  8. Closed Frequent Sequence Mining
  9. Huffman Encoding
  10. Decoder
  11. Simulation Results and Discussion
  12. Conclusion

Figures (3)

  • Figure 1: Huffman Tree for the Closed Frequent Sequences
  • Figure 2: Variation of $C_r$, PSNR, $k$ and $\alpha$ for Lena
  • Figure 3: (a) $\alpha$ vs Compression Time for Lena and Peppers (b) Lena Input Image (c), (d) and (e) are decompressed images at $k=8$, $k=15$ and $k=24$ respectively at $\alpha= 46\%$.