Optimasi Resolusi Citra pada MobileNetV3: Analisis Trade-Off Latensi dan Akurasi Deteksi Sampah
DOI:
https://doi.org/10.29407/6hnm2w65Abstract
Pengelolaan sampah di lingkungan sekolah memerlukan sistem pemilahan otomatis yang mampu beroperasi secara seketika. Namun, penerapan teknologi visi komputer pada perangkat standar sering terkendala oleh tingginya beban komputasi. Penelitian ini bertujuan mengimplementasikan arsitektur MobileNetV3 untuk mengklasifikasikan citra ke dalam kelas organik, anorganik, dan bukan sampah. Eksperimen difokuskan pada modifikasi dimensi citra masukan, yaitu ukuran 160, 224, dan 256 piksel, guna mengevaluasi keseimbangan antara ketepatan prediksi dan durasi komputasi. Hasil pengujian menunjukkan bahwa ukuran 160 piksel merupakan konfigurasi paling efisien yang mampu mencapai ketepatan tertinggi sebesar 0,9910 dengan waktu inferensi paling singkat, yakni 14,99 milidetik, tanpa dukungan pengolah grafis tambahan. Temuan empiris ini menegaskan kelayakan optimasi resolusi citra dalam mewujudkan infrastruktur deteksi sampah cerdas yang cepat, responsif, dan siap diterapkan pada perangkat berkapasitas rendah.
Keywords:
dimensi citra, klasifikasi sampah, MobileNetV3, visi komputer, waktu inferensi##plugins.themes.default.displayStats.downloads##
References
[1] M. Chhabra, B. Sharan, M. Elbarachi, and M. Kumar, ―Intelligent waste classification
approach based on improved multi-layered convolutional neural network,‖ pp. 84095–
84120, 2024, doi: 10.1007/s11042-024-18939-w.
[2] J. Teknologi, S. Informasi, J. Nurhidayah, Z. Zulfiandri, and S. Informasi, ―Tren dan
Tantangan Arsitektur Komputasi Neuromorfik : Tinjauan Literatur Sistematis,‖ vol. 6,
no. December 2024, pp. 103–113, 2025, doi: 10.35957/jtsi.v6i1.10046.
[3] M. A. Shafique and G. S. Member, ―ON-CNN : Low Latency and High Throughput
Online Arithmetic-Based Convolutional Neural Network Accelerator,‖ IEEE Access,
vol. 12, no. November, pp. 175698–175714, 2024, doi:
10.1109/ACCESS.2024.3502665.
[4] J. Zhong, J. Chen, A. Mian, and S. Member, ―DualConv : Dual Convolutional Kernels
for Lightweight Deep Neural Networks,‖ IEEE Trans. Neural Networks Learn. Syst.,
vol. 34, no. 11, pp. 9528–9535, 2023, doi: 10.1109/TNNLS.2022.3151138.
[5] J. Karim, O. F. Goni, M. Ahsan, J. Haider, and M. Kowalski, ―Enhancing agriculture
through real ‑ time grape leaf disease classification via an edge device with a
lightweight CNN architecture and Grad ‑ CAM,‖ Sci. Rep., pp. 1–23, 2024, doi:
10.1038/s41598-024-66989-9.
[6] J. Patel, H. Talsania, and K. Modi, ―Performance Evaluation of Indian Food Image
Classification system using Transfer Learning with MobileNetV3,‖ vol. 14, no. 3, 2022,
doi: 10.18090/samriddhi.v14i03.24.
[7] S. M. Raza, S. Murtaza, H. Abidi, and S. Y. Shin, ―DSConvNet : A Lightweight
Architecture for Extracting Image Features From Depthwise Separable Convolution
Network for Edge Devices,‖ IEEE Access, vol. 13, no. December, pp. 210102–210116,
2025, doi: 10.1109/ACCESS.2025.3639410.
[8] S. Naveen and M. R. Kounte, ―Optimized Convolutional Neural Network at the IoT
edge for image detection using pruning and quantization,‖ pp. 5435–5455, 2025, doi:
10.1007/s11042-024-20523-1.
[9] R. Nurul et al., ―WARNA DAN TEKSTUR MENGGUNAKAN ARTIFICAL
NEURAL NETWORK WASTE TYPE CLASSIFICATION SYSTEM BASED ON A
COMBINATION OF COLOR AND TEXTURE FEATURES USING ARTIFICIAL
NEURAL NETWORK,‖ vol. 11, no. 2, pp. 411–420, 2024, doi:
10.25126/jtiik.20241128330.
[10] L. Fischer-brandies, L. Müller, B. Rebholz, R. Buettner, and S. Member, ―To Combine
or Not to Combine ? The Influence of Combining Training Datasets on the Robustness
of Deep Learning Models : An Analysis for Optical Character Recognition of
Handwriting,‖ no. April, pp. 59039–59056, 2025, doi:
10.1109/ACCESS.2025.3556582.
[11] U. Sumalatha, K. K. Prakasha, S. Prabhu, and V. C. Nayak, ―Enhancing Finger Vein
Recognition With Image Preprocessing Techniques and Deep Learning Models,‖ IEEE
Access, vol. 12, no. November, pp. 173418–173440, 2024, doi:
10.1109/ACCESS.2024.3498601.
[12] L. Papa, P. Russo, I. Amerini, L. Zhou, and S. Member, ―A Survey on Efficient Vision
Transformers : Algorithms , Techniques , and Performance Benchmarking,‖ IEEE
Trans. Pattern Anal. Mach. Intell., vol. 46, no. 12, pp. 7682–7700, 2024, doi:
10.1109/TPAMI.2024.3392941.
[13] H. Zhu, Y. Liu, X. Gao, and L. Zhang, ―Combined CNN and Pixel Feature Image for
Fatty Liver Ultrasound Image Classification,‖ vol. 2022, 2022, doi:
10.1155/2022/9385734.
[14] M. H. Saleem, K. K. Velayudhan, and J. Potgieter, ―Weed Identification by SingleStage and Two-Stage Neural Networks : A Study on the Impact of Image Resizers and
Weights Optimization Algorithms,‖ vol. 13, no. April, pp. 1–19, 2022, doi:
10.3389/fpls.2022.850666.
[15] K. Riehl, M. Neunteufel, and M. Hemberg, ―Hierarchical confusion matrix for
classification performance evaluation,‖ J. R. Stat. Soc. Ser. C Appl. Stat., vol. 72, no. 5,
pp. 1394–1412, 2023, doi: 10.1093/jrsssc/qlad057.
[16] S. Amaliah and M. Nusrang, ―Penerapan Metode Random Forest Untuk Klasifikasi
Varian Minuman Kopi Di Kedai Kopi Konijiwa Bantaeng,‖ vol. 4, no. 2, pp. 121–127,
2022, doi: 10.35580/variansiunm31.
[17] O. Rainio, ―Evaluation metrics and statistical tests for machine learning,‖ Sci. Rep., pp.
1–14, 2024, doi: 10.1038/s41598-024-56706-x.
[18] M. Te Wu, ―Confusion matrix and minimum cross ‑ entropy metrics based motion
recognition system in the classroom,‖ Sci. Rep., no. 1821, pp. 1–10, 2022, doi:
10.1038/s41598-022-07137-z.
[19] C. Y. Kim, ―A novel MobileNet with selective depth multiplier to compromise
complexity and accuracy,‖ vol. 45, no. August 2022, pp. 666–677, 2023, doi:
10.4218/etrij.2022-0103.
[20] W. Lim and M. B. Bonab, ―An Aggressively Pruned CNN Model With Visual
Attention for Near Real-Time Wood Defects Detection on Embedded Processors,‖
IEEE Access, vol. 11, no. April, pp. 36834–36848, 2023, doi:
10.1109/ACCESS.2023.3266737.
[21] R. Islam et al., ―Deep Learning and Computer Vision Techniques for Enhanced
Quality Control in Manufacturing Processes,‖ IEEE Access, vol. 12, no. August, pp.
121449–121479, 2024, doi: 10.1109/ACCESS.2024.3453664.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Copyright on any article is retained by the author(s).
- The author grants the journal, right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work’s authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal’s published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
- The article and any associated published material is distributed under the Creative Commons Attribution-ShareAlike 4.0 International License
