Date of Award

2-29-2024

Publication Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

Keywords

Binary Images;DCT;Image Compression

Supervisor

Mohammad Hassanzadeh

Supervisor

Behnam Shahrrava

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

In an era where digital image content is increasing rapidly, the demand for efficient compression algorithms, especially those ensuring lossless quality, is more critical than ever. This is particularly true in the case of binary images that, despite their apparent simplicity, play a vital role in various applications, such as text documents, barcodes, and QR codes. Binary images are simple because they comprise only two colours and are represented by 1 bit per pixel. In this thesis, we introduce a novel lossless compression algorithm for binary images, employing a transform-based approach. The core innovation of this algorithm lies in using a distinctive property of a famous mathematical transform, the Discrete Cosine Transform (DCT), particularly DCT - Type III / DCT-III. The property enables the representation of a prime-length binary vector or one-dimensional array using a single DCT-III transformed coefficient. The development of this algorithm involves integrating established methodologies in compression to maximize its efficiency. Comparative testing against industry standards reveals that the algorithm achieves compression ratios on par with standards such as JBIG and outperforms algorithms like JPEG, JPEG-XR, and JPEG-LS. Additionally, in the thesis, we also explore the cascading of the newly proposed algorithm’s framework with existing compression techniques. This integration enhances the compression efficiency of these methods, demonstrating the algorithm’s versatility.

Download

Available for download on Friday, February 28, 2025

Included in

Engineering Commons

Share

COinS