Ornamental Plants Classification Using  Integration of Convolution With Capsule Network

MOHAMMAD FATONI; ERNASTUTI ERNASTUTI

doi:10.26760/mindjournal.v8i2.158-172

Ornamental Plants Classification Using Integration of Convolution With Capsule Network

MOHAMMAD FATONI, ERNASTUTI ERNASTUTI

Sari

Abstrak

Klasifikasi tanaman hias bertujuan untuk mempermudah media sosial tanaman hias dalam mengategorikan citra, sehingga sistem dapat merekomendasikan konten sesuai dengan preferensi pengguna. Pengguna berpotensi merasa cepat bosan apabila konten hanya ditampilkan secara acak. Penelitian ini melakukan Integration of Convolution with Capsule Network (ICCN) dengan menggabungkan beberapa lapisan strided convolution dan Capsule Network (CapsNet) untuk menghasilkan model klasifikasi yang memiliki komputasi lebih rendah dibandingkan original CapsNet dan mampu mengatasi permasalahan invariant of translation pada Convolutional Neural Network (CNN). Sebanyak 3 lapisan convolution dengan kernel berukuran 3x3 dan stride 2 ditambahkan pada CapsNet untuk membantu mengekstraksi citra dan mengurangi jumlah parameter yang dilatih. Hasil penelitian menunjukkan ICCN yang diusulkan memiliki jumlah parameter 2 kali lebih sedikit daripada original CapsNet dan memiliki akurasi lebih tinggi dibandingkan dengan CNN yaitu sebesar 95% sementara CNN berakurasi 93%.

Kata kunci: tanaman hias, klasifikasi citra, cnn, capsnet, integrasi

Abstract

The aim of ornamental plant classification is to assist ornamental plant social media in categorizing images, so the system is able to recommend content based on user preferences. Showing content randomly can lead to user boredom. This research implements Integration of Convolution with Capsule Network (ICCN) by combining several layers of strided convolution with Capsule Network (CapsNet) to create a classification model that has lower computation compared to the original CapsNet and able to address the issue of invariant of translation in Convolutional Neural Network (CNN). There are 3 convolutional layers with 3x3 kernel and stride of 2 added to CapsNet to assist in image extraction and reduce the number of trainable parameters. The research results showed that the proposed ICCN has 2 times fewer trainable parameters than the original CapsNet and achieves higher accuracy than CNN, with 95% accuracy, while CNN has an accuracy of 93%.

Keywords: ornamental plants, image classification, cnn, capsnet, integration

Teks Lengkap:

PDF

Referensi

Aggarwal, V. (2018). A Review: Deep Learning Technique for Image Classification. ACCENTS Transactions on Image Processing and Computer Vision, 4(11), 21 - 25.

Ayachi, R., Afif, M., Said, Y., & Atri, M. (2020). Strided Convolution Instead of Max Pooling for Memory Efficiency of Convolutional Neural Networks. International Conference on Sciences of Electronics, (pp. 234-243).

Ali, M., Hassan, M., Kifayat, K., Kim, J. Y., Hakak, S., & Khan, M. K. (2023). Social Media Content Classification and Community Detection Using Deep Learning and Graph Analytics. Technological Forecasting and Social Change, 188(1), 1-13.

Biscione, V., & Bowers, J. S. (2021). Convolutional Neural Networks Are Not Invariant to Translation, but They Can Learn to Be. The Journal of Machine Learning Research, 22(1), 10407-10434.

Eg, R., TÃ¸nnesen, Ã–. D., & Tennfjord, M. K. (2023). A Scoping Review of Personalized User Experiences on Social Media: The Interplay Between Algorithms and Human Factors. Computers in Human Behavior Reports, 9(1), 1-17.

Hutmacher, F., & Appel, M. (2023). The Psychology of Personalization in Digital Environments: From Motivation to Well-Beingâ€“a Theoretical Integration. Review of General Psychology, 27(1), 26-40.

Kauderer-Abrams, E. (2017, December 10). Quantifying Translation-Invariance in Convolutional Neural Networks. Retrieved from arxiv.org.

Mazzia, V., Salvetti, F., & Chiaberge, M. (2021). Efficient-CapsNet: Capsule Network With Self-Attention Routing. Scientific Reports, 11(1), 1-13.

Mittal, A., Kumar, D., Mittal, M., Saba, T., Abunadi, I., Rehman, A., & Roy, S. (2020). Detecting Pneumonia Using Convolutions and Dynamic Capsule Routing for Chest X-Ray Images. Sensors, 20(4), 1-30.

O'Shea, K., & Nash, R. (2015, November 26). An Introduction to Convolutional Neural Networks. Retrieved from arxiv.org.

Powers, D. M. (2020, October 11). Evaluation: From Precision, Recall and F-Measure to Roc, Informedness, Markedness and Correlation. Retrieved from arxiv.org.

Sabour, S., Frosst, N., & Hinton, G. E. (2017). Dynamic Routing Between Capsules. Neural Information Processing Systems, (pp. 1-11).

Sathya, R., & Abraham, A. (2013). Comparison of Supervised and Unsupervised Learning Algorithms for Pattern Classification. International Journal of Advanced Research in Artificial Intelligence, 2(2), 34-38.

Sharma, P., Arya, R., Verma, R., & Verma, B. (2023). Conv-CapsNet: Capsule Based Network for COVID-19 Detection Through X-Ray Scans. Multimedia Tools and Applications, 82(18), 28521â€“28545.

Singh, C. K., Gangwar, V. K., Majumder, A., Kumar, S., Ambwani, P. C., & Sinha, R. (2020). A Light-Weight Deep Feature Based Capsule Network. International Joint Conference on Neural Networks, (pp. 1-8).

Setyawan, D. (2022). Tinjauan Peningkatan Penjualan Tanaman Hias Di Masa Pandemi Dengan Life Cycle Assesment (LCA). National Multidisciplinary Sciences, (pp. 185-193).

Shah, P. L., Gupta, T. K., Dhakad, J. B., & Dâ€™silva, M. R. (2018). A Review Paper on Understanding Capsule Networks. International Journal of Engineering Development and Research, 6(4), 58-65.

Springenberg, J. T., Dosovitskiy, A., Brox, T., & Riedmiller, M. (2014, December 21). Striving for Simplicity: The All Convolutional Net. Retrieved from arxiv.org.

Tiwari, S., & Jain, A. (2021). Convolutional Capsule Network for COVIDâ€19 Detection Using Radiography Images. International Journal of Imaging Systems and Technology, 31(2), 525-539.

Wang, P., Fan, E., & Wang, P. (2021). Comparative Analysis of Image Classification Algorithms Based on Traditional Machine Learning and Deep Learning. Pattern Recognition Letters, 141(1), 61-67.

Yao, H., Tan, Y., Xu, C., Yu, J., & Bai, X. (2021). Deep Capsule Network for Recognition and Separation of Fully Overlapping Handwritten Digits. Computers & Electrical Engineering, 91(1), 1-12.

Yu, Z., Wang, K., Wan, Z., Xie, S., & Lv, Z. (2023). Popular Deep Learning Algorithms for Disease Prediction: A Review. Cluster Computing, 26(2), 1231-1251.

DOI: https://doi.org/10.26760/mindjournal.v8i2.158-172