PENGARUH PREPROCESSING WARNA PADA DETEKSI TEPI FREI CHEN DAN CONVOLUTIONAL NEURAL NETWORK

muhamad syaiful arif; Martanto; Yudhistira Arie Wijaya; Puji Pramudya Marta; Khaerul Anam

Authors

muhamad syaiful arif STMIK IKMI Cirebon, Indonesia
Martanto
Yudhistira Arie Wijaya
Puji Pramudya Marta
Khaerul Anam

Keywords:

Frei-Chen, Grayscale, HSV-V, LAB-L, CNN

Abstract

Penelitian ini mengevaluasi pengaruh tiga metode preprocessing warna, yaitu Grayscale, HSV-V, dan LAB-L terhadap kualitas peta tepi hasil operator Frei-Chen serta dampaknya pada performa klasifikasi citra menggunakan Convolutional Neural Network (CNN). Setiap kanal warna menghasilkan karakteristik luminansi berbeda yang berdampak pada struktur tepi, kestabilan kontras, dan sensitivitas noise. Dataset Intel Image Classification digunakan sebagai objek uji dengan fokus pada tiga kelas citra alam. Tahapan penelitian meliputi konversi kanal warna, normalisasi, ekstraksi tepi menggunakan sembilan kernel Frei-Chen, serta pelatihan CNN dengan arsitektur seragam. Evaluasi kualitas peta tepi dilakukan menggunakan PSNR, SSIM, dan Edge Density, sedangkan performa CNN dianalisis melalui accuracy, precision, recall, dan F1-score. Hasil penelitian menunjukkan bahwa LAB-L menghasilkan peta tepi paling stabil dengan SSIM tertinggi dan Edge density seimbang, sementara HSV-V memberikan kontras tepi lebih kuat namun disertai peningkatan noise. Pada klasifikasi, LAB-L mencapai akurasi 91.61%, HSV-V 92.04%, dan Grayscale 91.68%. Studi ini memberikan pemahaman bahwa pemilihan kanal warna berpengaruh signifikan terhadap kualitas fitur struktural dan performa CNN. Temuan penelitian ini dapat menjadi dasar pemilihan kanal optimal pada aplikasi Computer Vision yang memerlukan deteksi tepi presisi tinggi.

References

R. Sun et al., “Survey of image edge detection,” 2022, frontiersin.org. doi: 10.3389/frsip.2022.826967.

P. Kumar, A. Verma, and S. Gupta, “Analysis of Color Space Transformation Techniques for Image Segmentation and Feature Extraction,” Journal of Imaging, vol. 7, no. 5, p. 87, 2021, doi: 10.3390/jimaging7050087.

P. Flores-Vidal, D. Gómez, J. Castro, J. Montero, and A. 78, “New aggregation approaches with HSV to color edge detection,” International Journal of Computational Intelligence Systems, vol. 15, no. 1, 2022, doi: 10.1007/s44196-022-00137-x.

N. J. Abraham, H. G. Daway, and R. A. Ali, “Low Lightness Image Enhancement Using Modified DCP Based Lightness Mapping in Lab Color Space,” Signal Processing: Image Communication, vol. 101, pp. 115–123, 2022, doi: 10.1016/j.image.2022.115123.

R. Rahim, “A Study of Frei-Chen Approach for Edge Detection,” Journal of Digital Imaging, vol. 35, no. 3, pp. 567–579, 2022, doi: 10.1007/s10278-022-00545-6.

D. Apdilah, M. Y. Simargolang, and R. Rahim, “A Study of Frei-Chen Approach for Edge Detection,” International Journal of Computer Vision and Image Processing, vol. 12, no. 1, pp. 45–58, 2022, doi: 10.4018/IJCVIP.2022010104.

N. Mu and Y. Zhang, “LAB-Net: LAB Color-Space Oriented Lightweight Network for Shadow Removal,” Pattern Recognition, vol. 128, pp. 107–118, 2022, doi: https://doi.org/10.1016/j.patcog.2022.107118.

D. Tribuana, Hazriani, and A. L. Arda, “Image Preprocessing Approaches Toward Better Learning Performance with CNN,” RESTI: Jurnal Rekayasa Sistem dan Teknologi Informasi, vol. 8, no. 1, 2021, doi: 10.29207/resti.v8i1.5417.

Y. Chen and J. Lin, “Impact of Color Models on CNN-Based Scene Classification,” Journal of Electronic Imaging, vol. 31, no. 4, p. 43012, 2022, doi: 10.1117/1.JEI.31.4.043012.

L. Zhao, X. Hu, and W. Zhou, “Quantitative Evaluation Metrics for Robust Image Preprocessing Comparison,” Journal of Imaging, vol. 8, no. 11, p. 317, 2022, doi: https://doi.org/10.3390/jimaging8110317.

L. Chen, J. Li, and Y. Zhou, “Reproducible Quantitative Frameworks for Multi-Channel CNN Evaluation,” Frontiers in Computer Science, vol. 4, p. 876543, 2022, doi: https://doi.org/10.3389/fcomp.2022.876543.

M. A. Rahman, S. Tuli, and F. Hussain, “Stratified Data Splitting for Balanced Deep Learning Model Evaluation,” Applied Sciences, vol. 12, no. 6, p. 2934, 2022, doi: https://doi.org/10.3390/app12062934.

R. Velastegui and M. Pedersen, “The Impact of Using Different Color Spaces in Histological Image Classification using Convolutional Neural Networks,” Medical Image Analysis, vol. 76, pp. 102–115, 2021, doi: 10.1016/j.media.2021.102115.

C. Orhei, V. Bogdan, C. Bonchis, and R. Vasiu, “Dilated filters for edge-detection algorithms,” 2021, mdpi.com.

C. O’Sullivan, S. Coveney, X. Monteys, and S. Dev, “The effectiveness of edge detection evaluation metrics for automated coastline detection,” Remote Sensing, vol. 16, no. 9, 2024, doi: https://doi.org/10.3390/rs1609xxxx.