Claim Missing Document
Check
Articles

Found 5 Documents
Search
Journal : JURNAL TEKNIK INFORMATIKA DAN SISTEM INFORMASI

Implementasi Random Forest Untuk Klasifikasi Motif Songket Palembang Berdasarkan SIFT Siska Devella; Yohannes Yohannes; Firda Novia Rahmawati
JATISI (Jurnal Teknik Informatika dan Sistem Informasi) Vol 7 No 2 (2020): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)
Publisher : Lembaga Penelitian dan Pengabdian pada Masyarakat (LPPM) STMIK Global Informatika MDP

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35957/jatisi.v7i2.289

Abstract

Indonesia has a variety of intangible cultural heritage, one of which is songket. Songket has a lot of variety according to the characteristics of each region, especially Songket Palembang. Songket Palembang has more features compared to songket from other regions. Besides having historical value, Songket Palembang has a high motive, quality, and complexity in the manufacturing process. In this study, the Random Forest method was used to classify the Songket Palembang motif image by using Scale-Invariant Feature Transform (SIFT) feature extraction. The process of feature formation using the SIFT method is through the stages of extrema detection scale space, keypoint localization, orientation assignment, and keypoint descriptor. The resulting feature is used for the Random Forest classification. Songket motif images used in this study were 115 images of each type of motif, namely Chinese Flowers, Beautiful Flowers, and Pulir. Image selection is taken from 5 colors of each Songket Palembang motif. Training data and test data used were 100 and 15 for each Songket Palembang motif, respectively. The test results show that the SIFT and Random Forest methods for the classification of Songket Palembang motifs can provide a pretty good accuracy, where the SIFT and Random Forest methods can produce an overall accuracy of 92.98%, per class accuracy of 94.07%, precision 92.98%, and recall 89.74%.
Pengenalan ASL Menggunakan Metode Ekstraksi HOG dan Klasifikasi Random Forest Ningrum Larasati; Siska Devella; Muhammad Ezar Al Rivan
JATISI (Jurnal Teknik Informatika dan Sistem Informasi) Vol 8 No 2 (2021): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)
Publisher : Lembaga Penelitian dan Pengabdian pada Masyarakat (LPPM) STMIK Global Informatika MDP

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35957/jatisi.v8i2.456

Abstract

Sign languages ​​have many types, one of them is the American Sign Language (ASL). This study uses the ASL alphabet handshape image extracted with the Histogram of Oriented Gradient (HOG) feature and the resulting feature is used for the Random Forest classification. The test results show that using the HOG feature and the Random Forest classification method for ASL recognition gives a good accuracy rate, with an overall accuracy value of 99.10%, an average accuracy value per class of 77.43%, an average value of precision 88.81%, and an average recall value of 88.65%.
Penggunaan Fitur Saliency-SURF untuk Klasifikasi Citra Sel Darah Putih dengan Metode SVM Siska Devella; Yohannes Yohannes; Celvine Adi Putra
JATISI (Jurnal Teknik Informatika dan Sistem Informasi) Vol 8 No 4 (2021): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)
Publisher : Lembaga Penelitian dan Pengabdian pada Masyarakat (LPPM) STMIK Global Informatika MDP

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35957/jatisi.v8i4.1547

Abstract

Sel darah putih merupakan sel pembentuk komponen darah yang berfungsi melawan berbagai penyakit dari dalam tubuh (sistem kekebalan tubuh). Sel darah putih dibagi menjadi lima jenis, yaitu basofil, eosinofil, neutrofil, limfosit, dan monosit. Pendeteksian jenis sel darah putih dilakukan di laboratorium yang memerlukan seorang spesialis serta usaha yang lebih, waktu, dan biaya. Solusi yang dapat dilakukan salah satunya adalah menggunakan machine learning seperti support vector machine (SVM) dengan ekstraksi fitur SURF. Penelitian ini menggunakan dataset citra sel darah putih yang sebelumnya dilakukan tahap pre-processing yang, terdiri dari crop, resize, dan saliency. Metode saliency mampu memberikan bagian yang bermakna pada sebuah citra. Metode ekstraksi fitur SURF mampu memberikan keypoint yang dapat digunakan SVM dalam mengenali jenis sel darah putih. Penggunaan region-contrast saliency dengan kernel radial basis function (RBF) mendapatkan hasil akurasi, presisi, dan recall yang baik di bandingkan dengan penggunaan kernel lain dalam penelitian ini. Berdasarkan hasil pengujian yang didapat pada penelitian ini, saliency dapat meningkatkan hasil akurasi, presisi, dan recall dari SVM untuk dataset citra sel darah putih dibandingkan dengan tanpa saliency.
Klasifikasi Daun Herbal Berdasarkan Fitur Bentuk dan Tekstur Menggunakan KNN Meiriyama Meiriyama; Siska Devella; Sandra Mareza Adelfi
JATISI (Jurnal Teknik Informatika dan Sistem Informasi) Vol 9 No 3 (2022): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)
Publisher : Lembaga Penelitian dan Pengabdian pada Masyarakat (LPPM) STMIK Global Informatika MDP

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35957/jatisi.v9i3.2974

Abstract

Indonesia has an abundance of biodiversity. From a total of 40,000 types of herbal plants known in the world, there are approximately 30,000 types of herbal plants in Indonesia. Herbal plants are plants that are commonly used by people, especially in Indonesia, which have biodiversity as ingredients for making herbal medicines. Herbal plants are certainly not easy to recognize even though they often grow around the environment. Because there is still a lack of community knowledge about herbal plants, it is not possible to use these herbal plants. This study aims to classify the leaves of herbal plants using the K-Nearest Neighbor (KNN) method with k value is 3 and feature extraction of Histogram of Oriented Gradient (HOG) and Local Binary Patterns (LBP). The research was conducted on 15 types of herbal plants. Accuracy HOG method with KNN is 92.67%, Accuracy LBP with KNN is 88.67% and accuracy combination of HOG and LBP features with KNN method is 92.67%. Based on the three experiment scenarios that have been carried out, it shows that the combination of HOG and LBP features does not affect the accuracy of leaf classification of herbal plants.
Ekstraksi Fitur Warna dengan Histogram HSV untuk Klasifikasi Motif Songket Palembang Yohannes, Yohannes; Al Rivan, Muhammad Ezar; Devella, Siska; Meiriyama, Meiriyama
JATISI Vol 11 No 2 (2024): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)
Publisher : Universitas Multi Data Palembang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35957/jatisi.v11i2.8110

Abstract

Palembang Songket is a type of traditional woven cloth that has been registered as Indonesia's intangible cultural heritage since 2013. Palembang Songket has many motifs including Bunga Cina, Cantik Manis, and Pulir. The motifs on Palembang Songket have different meanings which can influence the selling price of the Songket. Recognition and classification of Palembang Songket types and motifs can be done by utilizing computer technology such as digital image processing and machine learning. In this research, the classification of Palembang Songket motifs was carried out using color features with histograms in Hue, Saturation, and Value (HSV) space and the Support Vector Machine (SVM) machine learning algorithm. Testing was carried out on a classification system using 45 test images. The histogram of HSV and SVM methods with the best kernel, namely RBF, were able to classify Palembang Songket motifs with an accuracy of 0.956; precision of 0.94; recall of 0.933; and f1-score of 0.931.