A stable and fast PSO algorithm guided by SPSA for vector quantization-based image compression

Abstract

Image compression plays a crucial role in reducing storage requirements and improving transmission efficiency. The effectiveness of lossy image compression using vector quantization (VQ) heavily depends on the quality of codebook generation, which is inherently an optimization problem. In this paper, a coupled hybrid algorithm integrating Simultaneous Perturbation Stochastic Approximation (SPSA) into Particle Swarm Optimization (PSO) is proposed to enhance both the convergence speed and codebook quality in vector quantization. The novel SPSA-FPSO algorithm, by generating multiple alternative codebooks at each iteration and selecting the best, successfully avoids local minima and achieves faster convergence. Experimental results, conducted on standard gray-level images of various contrast levels, demonstrate that the proposed SPSA-FPSO algorithm outperforms both basic PSO and SPSA algorithms in terms of lower mean square error (MSE) and higher convergence speeds, establishing its superiority for VQ-based image compression tasks. This superiority is also shown to be valid when compared to other metaheuristic algorithms.

Keywords

• Image compression
• Metaheuristic algorithms
• Vector Quantization
• Codebook generation
• PSO
• SPSA

Vector nicemleme tabanlı görüntü sıkıştırma için EPSY ile yönlendirilen kararlı ve hızlı bir PSO algoritması

Öz

Görüntü sıkıştırma, depolama gereksinimlerini azaltmak ve iletim verimliliğini artırmak açısından büyük bir öneme sahiptir. Vektör nicemleme (VN) tabanlı kayıplı görüntü sıkıştırmanın başarısı, esasen bir optimizasyon problemi olan kod tablosu üretiminin kalitesine bağlıdır. Bu makalede, hem algoritmanın yakınsama hızını hem de VN kod tablosunun kalitesini artırmak için Eşzamanlı Pertürbasyon Stokastik Yaklaşımı (EPSY) tekniğini Parçacık Sürü Optimizasyonu (PSO) ile bütünleştiren hibrit bir algoritma önerilmektedir. Önerilen EPSY-HPSO algoritması, her iterasyonda birden fazla alternatif kod kitabı üreterek en iyisini seçmekte ve yerel minimum noktalarından kaçınarak daha hızlı bir yakınsama sağlamaktadır. Farklı kontrast seviyelerine sahip standart gri seviye görüntüler üzerinde gerçekleştirilen deneysel sonuçlar, EPSY-HPSO algoritmasının hem ortalama kare hata (OKH) değerlerini düşürme hem de daha yüksek yakınsama hızları açısından klasik PSO ve EPSY algoritmalarından daha başarılı olduğunu göstererek VN tabanlı görüntü sıkıştırmadaki üstünlüğünü kanıtlamaktadır. Bu üstünlüğün diğer metasezgisel algoritmalarla karşılaştırıldığında da geçerli olduğu gösterilmektedir

Anahtar Kelimeler

• Görüntü sıkıştırma
• Metasezgisel algoritmalar
• Vektör nicemleme
• Kod tablosu üretimi
• PSO
• EPSY

Kaynakça

1. [1] Gray RM. “Vector Quantization”. IEEE ASSP Magazine, 1(1), 4-29, 1984.
2. [2] Wu Z, Yu J. “Vector quantization: a review”. Frontiers of Information Technology & Electronic Engineering, 20(4), 507-524, 2019.
3. [3] Kumar G, Kumar R. “Analysis of Arithmetic and Huffman Compression Techniques by Using DWT-DCT.” International Journal of Image, Graphics and Signal Processing, 4, 63-70, 2021.
4. [4] Lu TC, Chang CY. “A Survey of VQ Codebook Generation”. Journal of Information Hiding and Multimedia Signal Processing, 1(3), 190-203, 2010.
5. [5] Yang SB. “Constrained-Storage multistage vector quantization based on genetic algorithms”. Pattern Recognition, 41(2), 689–700, 2008.
6. [6] Chiranjeevi K, Jena UR. “Image compression based on vector quantization using cuckoo search optimization technique”. Ain Shams Engineering Journal, 9(4), 1417–1431, 2018.
7. [7] Tsai CW, Tseng SP, Yang CS, Chiang MC. “Preaco: A fast ant colony optimization for codebook generation”. Applied Soft Computing, 13(6), 3008–3020, 2013.
8. [8] Feng HM, Chen CY, Ye F. “Evolutionary fuzzy particle swarm optimization vector quantization learning scheme in image compression”. Expert Systems with Applications, 32(1), 213–222, 2007.

9. [9] Horng MH. “Vector quantization using the firefly algorithm for image compression”. Expert Systems with Applications, 39(1), 1078–1091, 2012.
10. [10] Guo JR, Wu CY, Huang ZL, Wang FJ, Huang MT. “Vector quantization image compression algorithm based on bat algorithm of adaptive separation search”. International Conference on Advanced Intelligent Systems and Informatics, Cairo, Egypt, 11-13 December 2021.
11. [11] Geetha K, Anitha V, Elhoseny M, Kathiresan S, Shamsolmoali P, Selim MM. “An evolutionary lion optimization algorithm-based image compression technique for biomedical applications”. Expert Systems, 38(1), Article e12508, 2021.
12. [12] Kumari GV, Rao GS, Rao BP. “Flower pollination-based K-means algorithm for medical image compression”. International Journal of Advanced Intelligent Paradigms, 18(2), 171–192, 2021
13. [13] Rahebi J. “Vector quantization using whale optimization algorithm for digital image compression”. Multimedia Tools and Applications, 81(14), 20077–20103, 2022.
14. [14] Althobaiti MM. Crow Search Algorithm Based Vector Quantization Approach for İmage Compression in 6G Enabled Industrial Internet of Things Environment. Editors: Gupta D, Ragab M, Mansour RF, Khamparia A, Khanna A. AI-enabled 6G networks and applications, 55–73, Wiley, 2023.
15. [15] Nag S. “Vector quantization using the improved differential evolution algorithm for image compression”. Genetic Programming and Evolvable Machines, 20, 187–212, 2019.
16. [16] Ghadami R, Rahebi J. “Compression of images with a mathematical approach based on sine and cosine equations and vector quantization (VQ)”. Soft Computing, 27(22), 17291–17311, 2023.
17. [17] Kilic I, Cetin M. “Improved adolescent identity search algorithm for block-based image compression”. Expert Systems with Applications, 237, 121715, 2024.
18. [18] Kilic I. “A novel codebook generation by smart fruit fly algorithm based on exponential flight”. The International Arab Journal of Information Technology, 20(4), 584-591, 2023.
19. [19] Kilic I. “A Levy flight based BAT optimization algorithm for block-based image compression”. Technicki Glasnik–Technical Journal, 16(4), 477-483, 2022.
20. [20] Spall JC. "Multivariate stochastic approximation using a simultaneous perturbation gradient approximation". IEEE Transactions on Automatic Control, 37(3), 332-341, 1992.
21. [21] Chen S, Mei T, Luo M, Yang X. “Identification of nonlinear system based on a new hybrid gradient-based PSO algorithm”. ICIA 2007 International Conference on Information Acquisition, Jeju City, Korea, 8-11 July 2007.
22. [22] Kaveh A, Talatahari S. “A hybrid particle swarm and ant colony optimization for design of truss structures”. Asian Journal of Civil Engineering, 9(4), 329–48, 2008.
23. [23] Plevris V, Papadrakakis M. “A hybrid particle swarm-gradient algorithm for global structural optimization”. Computer-Aided Civil and Infrastructure Engineering, 26(1), 48-68, 2011.
24. [24] Cherki I, Chaker A, Djidar Z, Khalfallah N, Benzergua F. “A sequential hybridization of genetic algorithm and particle swarm optimization for the optimal reactive power flow”. Sustainability, 11(14), 3862, 2019.
25. [25] Seyedpoor SM, Gholizadeh S. Talebian SR. “An efficient structural optimisation algorithm using a hybrid version of particle swarm optimisation with simultaneous perturbation stochastic approximation”. Civil Engineering and Environmental Systems, 27(4), 295–313, 2010.
26. [26] Wessels S, van der Haar D. “Using particle swarm optimization with gradient descent for parameter learning in convolutional neural networks”. CIARP 25th Iberoamerican Congress, Porto, Portugal, 10-13 May 2021.
27. [27] Pujari AK, Veeramachaneni SD. “Gradient based hybridization of PSO”. CSAI 2023 International Conference on Computer Science and Artificial Intelligence, Beijing, China, 08-10 December 2023.
28. [28] Barroso ES, Parente JE, Cartaxo de Melo AM. “A hybrid PSO-GA algorithm for optimization of laminated composites”. Structural and Multidisciplinary Optimization, 55(6), 2111–2130, 2017.
29. [29] Parouha RP, Verma P. “Design and applications of an advanced hybrid meta-heuristic algorithm for optimization problems”. Artificial Intelligence Review, 54, 5931–6010, 2021.
30. [30] Shankar T, Shanmugavel S, Rajesh A. “Hybrid HSA and PSO algorithm for energy efficient cluster head selection in wireless sensor networks”. Swarm and Evolutionary Computation, 30, 1–10, 2016.
31. [31] Chegini SN, Bagheri A, Najafi F. “PSOSCALF: A new hybrid PSO based on Sine and Cosine Algorithm and Levy Flight for solving optimization problems”. Applied Soft Computing, 73, 697-726, 2018.
32. [32] Şenel FA, Gokce F, Yuksel AS, Yigit T. “A novel hybrid PSO–GWO algorithm for optimization problems”. Engineering with Computers, 35(4), 1359–1373, 2019.
33. [33] Kaya S, Karaçi̇zmeli̇ İH, Aydilek İB, Tenekeci ME, Gümüşçü A. “The effects of initial populations in the solution of flow shop scheduling problems by hybrid firefly and particle swarm optimization algorithms”. Pamukkale University Journal of Engineering Science, 26(1), 140–149, 2020.
34. [34] Qiao J, Wang G, Yang Z, Luo X, Chen J, Li K, Li P. “A hybrid particle swarm optimization algorithm for solving engineering problem”. Scientific Reports, 14(1), 8357, 2024.
35. [35] Kiranyaz S, Ince T, Gabbouj M. “Stochastic approximation driven particle swarm optimization with simultaneous perturbation-who will guide the guide?”. Applied Soft Computing, 11(2), 2334–2347, 2011.
36. [36] Lloyd SP. “Least squares quantization in PCM”. IEEE Transactions on Information Theory, IT-28(2), 129-137, 1982.
37. [37] Pena JM, Lozano JA. Larranage P. “An empirical comparison of four initialization methods for the K-Means algorithm”. Pattern Recognition Letters, 20, 1027-1040, 1999.
38. [38] Kennedy J, Eberhart R. “Particle swarm optimization”. IEEE International Conference on Neural Networks, Australia, 27 November- 1 December 1995.
39. [39] Shi Y, Eberhart R. “A modified particle swarm optimizer”. IEEE Congress on Evolutionary Computation, USA; 4-9 May 1998.
40. [40] Gad AG. “Particle Swarm Optimization algorithm and its applications: A systematic review”. Archives of Computational Methods in Engineering, 29, 2531–2561, 2022.
41. [41] Spall JC. “Implementation of the simultaneous perturbation algorithm for stochastic optimization”. IEEE Transactions on Aerospace and Electronics Systems, 34(3), 817-823, 1998.