Pengembangan Sistem Dermatologi Veteriner Kucing dan Anjing Berbasis Two-Stage Lesion-Aware Pipeline dengan Mekanisme Confidence Calibration

Assariy Ramdhani, Randi Proska Sandra, Vikri Aulia, Dony Novaliendry

Abstract


Tingginya prevalensi penyakit kulit anjing dan kucing di Indonesia serta kemiripan visual antarkondisi sering memicu misdiagnosis. Mengatasi kelemahan detektor konvensional pada lesi kecil, penelitian ini mengembangkan Two-Stage Lesion-Aware Pipeline berbasis YOLOv12 untuk lokalisasi dan EfficientNetV2-S untuk klasifikasi sebagai alat penapisan dermatologi hewan terintegrasi AI. Sistem ini diperkuat dengan Temperature Scaling untuk kalibrasi kepercayaan model dan diimplementasikan pada aplikasi mobile berbasis Flutter/FastAPI. Hasil evaluasi menunjukkan performa [email protected] sebesar 0,89 (kucing) dan 0,64 (anjing), serta F1-score weighted klasifikasi mencapai 0,9976 dan 0,9871. Secara signifikan, kalibrasi (T=1,3174) berhasil mereduksi Expected Calibration Error (ECE) sebesar 32,4% menjadi 0,0209, memenuhi target fungsional (ECE<0,15) dengan rata-rata waktu respons sistem 4,97 detik. Evaluasi usabilitas mencatat skor Excellent (85), sementara penilaian pakar memvalidasi kualitas antarmuka (4,44/5,00). Pipeline terkalibrasi ini terbukti layak sebagai instrumen penapisan lini pertama, dengan rekomendasi perluasan cakupan penyakit dan validasi pada test set independen sebelum diadopsi secara klinis.

Keywords


dermatologi veteriner; Two-stage lesion-aware pipeline; YOLOv12; EfficientNetV2; confidence calibration.

Full Text:

PDF

References


Abudukelimu, H., Gao, Y., Abulizi, A., Musideke, M., Wu, S., Wang, M., Aizizi, M., Yehaiya, G., & Abudukelimu, M. (2025). DVF-YOLO-Seg: A two-stage breast mass segmentation model with enhanced feature extraction and small lesion detection. DIGITAL HEALTH, 11. https://doi.org/10.1177/20552076251374192

Azmoodeh-Kalati, M., Shabani, H., Maghareh, M. S., Barzegar, Z., & Lashgari, R. (2025). Leveraging an ensemble of EfficientNetV1 and EfficientNetV2 models for classification and interpretation of breast cancer histopathology images. Scientific Reports, 15(1). https://doi.org/10.1038/s41598-025-06853-6

Bard, A. M., Hinchliffe, S., Chan, K. W., Buller, H., & Reyher, K. K. (2023). ‘I Believe What I’m Saying More Than the Test’: The Complicated Place of Rapid, Point-of-Care Tests in Veterinary Diagnostic Practice. Antibiotics, 12(5). https://doi.org/10.3390/antibiotics12050804

Cao, Z., Li, Y., Kim, D. H., & Shin, B. S. (2024). Deep Neural Network Confidence Calibration from Stochastic Weight Averaging. Electronics (Switzerland), 13(3). https://doi.org/10.3390/electronics13030503

Dao, T., Fu, D. Y., Ermon, S., Rudra, A., & Ré, C. (2022). FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness. https://doi.org/10.48550/arXiv.2205.14135

Du, Y., Wang, Y., Kong, M., Liang, T., Long, Q., Chen, B., & Zhu, Q. (2025). Confidence Calibration for Multimodal LLMs: An Empirical Study through Medical VQA. https://doi.org/10.1007/978-3-032-04978-0_9

Kannaiya Raja, N., Venkata Naga Ramesh, J., Yousef ABaker El-Ebiary, T., Muniyandy, E., Konda Reddy, N., Ravi Kumar, V., Devarasetty, P., & Associate Professor, S. (2025). Pose Estimation of Spacecraft Using Dual Transformers and Efficient Bayesian Hyperparameter Optimization. In IJACSA) International Journal of Advanced Computer Science and Applications (Vol. 16, Number 4). https://doi.org/10.14569/IJACSA.2025.0160463

Karandikar, A., Cain, N., Tran, D., Lakshminarayanan, B., Shlens, J., Mozer, M. C., & Roelofs, B. (2021). Soft Calibration Objectives for Neural Networks. https://doi.org/10.48550/arXiv.2108.00106

Khanam, R., & Hussain, M. (2025). A Review of YOLOv12: Attention-Based Enhancements vs. Previous Versions. https://doi.org/10.48550/arXiv.2504.11995

Nirthika, R., Manivannan, S., Ramanan, A., & Wang, R. (2022). Pooling in convolutional neural networks for medical image analysis: a survey and an empirical study. In Neural Computing and Applications (Vol. 34, Number 7, pp. 5321–5347). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/s00521-022-06953-8

Qian, W., & Yi, Z. (2025). MedVisioneer: A Lesion-Aware Deep Vision Framework for Enhanced Diagnostic Assistance in CT Image Recognition and Adaptation. IEEE Access. https://doi.org/10.1109/ACCESS.2025.3648493

Tan, M., & Le, Q. V. (2021). EfficientNetV2: Smaller Models and Faster Training. https://doi.org/10.48550/arXiv.2104.00298

Tekin, H., Kılıç, Ş., & Doğan, Y. (2025). DiagNeXt: A Two-Stage Attention-Guided ConvNeXt Framework for Kidney Pathology Segmentation and Classification. Journal of Imaging, 11(12). https://doi.org/10.3390/jimaging11120433

Tian, Y., Ye, Q., & Doermann, D. (2025). YOLOv12: Attention-Centric Real-Time Object Detectors. http://arxiv.org/abs/2502.12524

Tomani, C., Cremers, D., & Buettner, F. (2022). Parameterized Temperature Scaling for Boosting the Expressive Power in Post-Hoc Uncertainty Calibration. https://doi.org/10.48550/arXiv.2102.12182

Yu, Y., Gomez-Cabello, C. A., Haider, S. A., Genovese, A., Prabha, S., Trabilsy, M., Collaco, B. G., Wood, N. G., Bagaria, S., Tao, C., & Forte, A. J. (2025). Enhancing Clinician Trust in AI Diagnostics: A Dynamic Framework for Confidence Calibration and Transparency. Diagnostics, 15(17). https://doi.org/10.3390/diagnostics15172204

Zaki, F. H. M., Airin, C. M., Nururrozi, A., Yanuartono, & Indarjulianto, S. (2021). A review: the prevalence of dermatophytosis on cats in Indonesia and Turkey. BIO Web of Conferences, 33. https://doi.org/10.1051/bioconf/20213306004

Zheng, Y., Qiu, Y., Che, H., Chen, H., Zheng, W.-S., & Wang, R. (2024). Deep Model Reference: Simple yet Effective Confidence Estimation for Image Classification. https://doi.org/10.1007/978-3-031-72117-5_17




DOI: https://doi.org/10.17509/ijdb.v6i1.103839

Refbacks

  • There are currently no refbacks.


Copyright (c) 2026 Universitas Pendidikan Indonesia (UPI)

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

Indonesian Journal of Digital Business is published by Universitas Pendidikan Indonesia (UPI)
and managed by Department of Digital Business
Jl. Dr. Setiabudi No.229, Kota Bandung, Indonesia - 40154
View My Stats