SEGMENTASI CITRA DIGITAL IKAN MENGGUNAKAN METODE THRESHOLDINGABSTRAK Untuk mengenal jenis-jenis ikan berdasarkan ciri-cirinya, telah dibuat suatu sistem untuk memisahkan objek mata ikan menggunakan metode thresholding. Prosesnya dimulai dengan menginput citra digital ikan, selanjutnya dikonversi ke citra grayscale. Kemudian dilakukan proses segmentasi terhadap citra grayscale. Selanjutnya, dipilih hasil segmentasi dan ditandai dengan proses deteksi tepi menggunakan operator Canny yang dipertajam dengan proses dilasi. Proses terakhir adalah membuat plot contour terhadap hasil proses dilasi dan citra grayscale. Hasil segmentasi berhasil memisahkan objek mata ikan dengan menggunakan metode thresholding local. Keseluruhan proses ini dilakukan dengan menggunakan MATLAB R2012a. Kata kunci : Mata Ikan, Segmentasi Citra, Thresholding DIGITAL FISH IMAGE SEGMENTATION BY THRESHOLDING METHOD ABSTRACT A system of fish eyelets seperation has been conducted to identify types of fish acording to their characteristics, by using thresholding method. The process start by inserting digital fish image then convert it to grayscale image. Next step is to process segmentation the grayscale image. Choosed the segmentation result then marked it by edge detection process using Canny operation which has been sharpened by dilation process. The last step is to make contour plot to dilation result and grayscale image. The result of the segmentation shows that the fish eyelets can be separated using local thresholding method. The whole process is conducted by using MATLAB R2012a. Keywords : Fish Eyelets, Segmentation Image, Thresholding
30 December 2022 in Bulletin of Social Informatics Theory and Application
Indonesia has a variety of medicinal plants that are efficacious for preventing or treating various diseases. Each region has unique medicinal plants, such as in North Sulawesi, there are many medicinal plants with local names of "Jarak" (Jatropha curcas), "Jarak Merah" (Jatropha multifida), "Miana" (Coleus Scutellarioide), and "Sesewanua" (Clerodendron Squmatum Vahl). This research applies the Convolutional Neural Network (CNN) method to identify the types of medicinal plants of North Sulawesi based on leaf images. Data was collected directly by taking photos of medicinal plant leaves and then using the augmentation process to increase the data. The first stage is conducting training and validation processes using 10-fold cross-validation, resulting in 10 classification models. Evaluation results show that the lowest validation accuracy of 98.4% was obtained from fold-4, and the highest was 100% from fold-2. The third stage was to run the testing process using new data. The results showed that the worst model produced a test accuracy of 80.91% while the best model produced an accuracy of 87.73% which means that the identification model is quite good and stable in classifying types of medicinal plants based on its leaf images. The final stage is to develop a web-based system to deploy the best model so people can use it in real-time
30 June 2022 in Knowledge Engineering and Data Science
The human facial expression identification system is essential in developing human interaction and technology. The development of Artificial Intelligence for monitoring human emotions can be helpful in the workplace. Commonly, there are six basic human expressions, namely anger, disgust, fear, happiness, sadness, and surprise, that the system can identify. This study aims to create a facial expression identification system based on basic human expressions using the Convolutional Neural Network (CNN) with a 16-layer VGG architecture. Two thousand one hundred thirty-seven facial expression images were selected from the FER2013, JAFFE, and MUG datasets. By implementing image augmentation and setting up the network parameters to Epoch of 100, the learning rate of 0,0001, and applying in the 5Fold Cross Validation, this system shows performance with an average accuracy of 84%. Results show that the model is suitable for identifying the basic facial expressions of humans
“Bidik Misi” (BM) is a scholarship from the Government of Indonesia with two criteria: economic need and academic performance. Main candidate targets are those from low economic families. Due to the large number of potential recipients who desire to continue their study to the university level despite the limited quota, difficulties occur in selecting candidates. This research aims to provide a new selection model by combining Back-Propagation Neural Network (BPNN) on Fuzzy Inferences System (FIS) for the BM scholarship selection process where BPNN is used as classifier to eliminate non-recommended candidates to reduce the system's workload before applying FIS as candidate selector. By considering spearman's correlation coefficients of the input parameters in creating fuzzy framework for BS scholarship selection, the accuracy of the system is superior to previous works. In the future will be useful in making an automated selection system in order to help the committee by reducing their manual labor.
Terumbu karang bermanfaat bagi kelangsungan hidup makhluk bawah air. Namun saat ini terancam keberadaannya, salah satu penyebab adalah peningkatan suhu Bumi sehingga banyak terumbu karang mengalami kerusakan ditandai dengan pemutihan warna karang. Pengamatan kesehatan karang dapat dilakukan dengan tabel kesehatan karang (Coral Health Chart). Penelitian ini dimulai dengan menghitung nilai rata-rata RGB (red, green, blue) dengan pengolahan citra digital terhadap Coral Health Chart, dan menerapkan metode kuadrat terkecil untuk memperoleh perumusan dalam rangka penentuan tingkat kesehatan dari terumbu karang menggunakan bantuan komputer. Langkah selanjutnya hitung nilai rata-rata RGB dari beberapa gambar karang dengan pengolahan citra digital, kemudian diterapkan pada perumusan yang telah diperoleh sebelumnya. Penelitian dilakukan pada 8 contoh gambar karang, hasilnya menunjukan tingkat kesehatan karang dan persentase tingkat kesehatan dari masing-masing karang yaitu untuk karang dengan simbol j=95.52%, k=65.02%, l=80.42%, m=81.11%, n=84.69%, o=52.87%, p=79.41% dan karang q=69.78%. Kata kunci : Tabel Kesehatan Karang, Pengolahan Citra Digital, Kesehatan Karang, Metode Kuadrat Terkecil
Nowadays, information technology especially machine learning has been used to evaluate the feasibility of debtors. One of the challenges in this classification model is the occurrence of imbalanced datasets, especially in the German Credit Dataset. Another challenge is developing an optimal model for evaluating debtor eligibility. Based on these challenges, this study aims to develop an optimal model for evaluating debtor eligibility on the German Credit Dataset, using the decision trees, k-Nearest Neighbor (k-NN) and Synthetic Minority Oversampling Technique (SMOTE). SMOTE and k-NN is used to overcome challenges regarding imbalanced datasets. While the decision tree are applied to produce a debtor classification model. In general, the steps taken are preparing datasets, pre-processing data, dividing datasets, oversampling with SMOTE, and classification models using decision trees, and testing. Model performance evaluation is represented by accuracy values obtained from the confusion matrix and area under curve (AUC) values generated by the Receiver Operating Characteristic (ROC). Based on the tests that have been carried out, the best accuracy value in the test is obtained at 73.00% and the AUC value is 0.708, in parameters k = 3 and Max-Depth = 25. Based on the analysis produced, the proposed model can improve performance compared to if the dataset is not applied SMOTE.