Claim Missing Document
Check
Articles

Found 36 Documents
Search

Identification of mangrove tree species using deep learning method Paranita Asnur; Rifki Kosasih; Sarifuddin Madenda; Dewi A. Rahayu
International Journal of Advances in Applied Sciences Vol 12, No 2: June 2023
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijaas.v12.i2.pp163-170

Abstract

Artificial intelligence can help classify plants to make identification easier for everyone. This technology can be used to classify mangrove trees. The degradation of mangrove forests has resulted in a 20% loss of biodiversity, an 80% loss of microbial decomposers, reduced C-organic soil, and fish spawning grounds, resulting in estimated losses in the ecological and economic fields for up to IDR 39 billion. The identification of different mangrove species is the first step in ensuring the preservation of these forests. Therefore, this research aimed to develop algorithms and a convolutional neural network (CNN) architecture to classify mangrove tree species with the highest possible accuracy using Python software. The architecture selection for this model includes a batch size of 32, an input image size of 128x128 pixels, four classes, four convolution layers, four rectified linear unit (ReLU) layers, 2x2 max-pooling, and two fully connected layers (FCL). The finding showed that the resulting accuracy from the test was 97.50%, while the validation test was 81.25%, applied to four types of mangrove leaves, including Avicenia marina, Avicenia officialis, Rizophora apiculata, and Soneratia caseolaris.
Multidisciplinary classification for Indonesian scientific articles abstract using pre-trained BERT model Antonius Angga Kurniawan; Sarifuddin Madenda; Setia Wirawan; Ruddy J. Suhatril
International Journal of Advances in Intelligent Informatics Vol 9, No 2 (2023): July 2023
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/ijain.v9i2.1051

Abstract

Scientific articles now have multidisciplinary content. These make it difficult for researchers to find out relevant information. Some submissions are irrelevant to the journal's discipline. Categorizing articles and assessing their relevance can aid researchers and journals. Existing research still focuses on single-category predictive outcomes. Therefore, this research takes a new approach by applying a multidisciplinary classification for Indonesian scientific article abstracts using a pre-trained BERT model, showing the relevance between each category in an abstract. The dataset used was 9,000 abstracts with 9 disciplinary categories. On the dataset, text preprocessing is performed. The classification model was built by combining the pre-trained BERT model with Artificial Neural Network. Fine-tuning the hyperparameters is done to determine the most optimal hyperparameter combination for the model. The hyperparameters consist of batch size, learning rate, number of epochs, and data ratio. The best hyperparameter combination is a learning rate of 1e-5, batch size 32, epochs 3, and data ratio 9:1, with a validation accuracy value of 90.8%. The confusion matrix results of the model are compared with the confusion matrix results by experts. In this case, the highest accuracy result obtained by the model is 99.56%. A software prototype used the most accurate model to classify new data, displaying the top two prediction probabilities and the dominant category. This research produces a model that can be used to solve Indonesian text classification-related problems.
ENKRIPSI CITRA DIGITAL MENGGUNAKAN KOMPOSISI TRANSPOSISI CAT MAP DAN SUBTITUSI KEY STREAM LOGISTIC MAP Rama Dian Syah; Sarifuddin Madenda; Ruddy J. Suhatril; Suryadi Harmanto
Jurnal Ilmiah Teknologi dan Rekayasa Vol 28, No 3 (2023)
Publisher : Universitas Gunadarma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35760/tr.2023.v28i3.7951

Abstract

Transmisi pertukaran data digital melalui jaringan internat menjadi hal penting pada kemajuan teknologi. Risiko pembajakan oleh pihak yang tidak bertanggung jawab mungkin terjadi karena kemudahan dalam pertukaran data. Pengembangan metode enkripsi data yang andal dan kuat adalah solusi untuk risiko ini. Penelitian ini mengusulkan algoritma enkripsi data baru melalui komposisi enkripsi transposisi Cat Map dan enkripsi substitusi Logistic Map. Algoritma yang diusulkan secara bersamaan mengubah posisi data dan mengubah nilai data secara acak. Penelitian telah dilakukan dengan menggunakan beberapa citra dengan berbagai fitur dan ukuran yang berbeda. Analisis keacakan citra hasil enkripsi menunjukkan bahwa histogram intensitas warna piksel memiliki distribusi yang seragam dengan nilai korelasi rendah mendekati 0. Hasil analisis peak signal to noise ratio (PSNR) menunjukkan citra hasil dekripsi sama dengan citra asli . Algoritma yang diusulkan memiliki ruang kunci 3.24 × 1068. Hasil NPCR, UACI dan Entropi menunjukkan algoritma yang diusulkan tahan terhadap serangan diferensial dan serangan entropi.
Deteksi Seksisme Online menggunakan Support Vector Machine dan Naïve Bayes SHABIRA, DIYANK; MADENDA, SARIFUDDIN; SIAGIAN, AL HAFIZ AKBAR MAULANA; RIYANTO, SLAMET
MIND (Multimedia Artificial Intelligent Networking Database) Journal Vol 8, No 2 (2023): MIND Journal
Publisher : Institut Teknologi Nasional, Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26760/mindjournal.v8i2.254-266

Abstract

AbstrakSeksisme online menjadi topik penting di media sosial yang mempengaruhi perkembangan internet, menimbulkan efek negatif dan menjadi ancaman serius bagi wanita yang menjadi target. Penelitian ini menggunakan machine learning untuk mendeteksi seksisme pada kalimat bahasa Inggris. Algoritma yang digunakan adalah Support Vector Machine dan Naive Bayes. Grid search diterapkan pada model untuk mencari kombinasi hyperparameter terbaik sehingga menghasilkan skor terbaik. Pelatihan dibagi menjadi dua tugas, yaitu (1) pelatihan model menggunakan data tanpa penanganan imbalanced dan (2) pelatihan model menggunakan data yang telah dilakukan SMOTE. Hasil dari pelatihan model menunjukkan model SVM+SMOTE menghasilkan rata-rata skor F1 terbaik paling tinggi yaitu sebesar 0,96. Pengujian menggunakan data uji menunjukkan model SVM+SMOTE menghasilkan skor F1 tertinggi, yaitu sebesar 0,90 dengan 1467 kalimat diklasifikasikan benar 'not sexist’, 47 kalimat ‘not sexist’ diklasifikasikan sebagai ‘sexist’, 189 kalimat ‘sexist’ diklasifikasikan benar dan 297 kalimat ‘sexist’ diklasifikasikan sebagai ‘not sexist’.Kata kunci: Seksisme, Deteksi, SVM, Naive Bayes, SMOTEAbstractOnline sexism has become a significant issue on social media, impacting internet progress and posing a serious threat to targeted women. This research uses machine learning to detect sexism in English sentences. The algorithms used are Support Vector Machine and Naive Bayes. Grid search is applied in the model to find the best combination of hyperparameters to produce the best score. The training is divided into two tasks: (1) training the model using unhandle the imbalanced data and (2) training the model using data with SMOTE. The training results show that the SVM+SMOTE model produces the highest average best F1 score is 0.96. The testing results show that the SVM+SMOTE model produces the highest F1 score is 0.90 with 1467 sentences correctly classified as 'not sexist', 47 'not sexist' sentences classified as 'sexist', 189 sentences classified as 'sexist' correctly and 297 'sexist' sentences were classified as 'not sexist'.Keywords: Sexism, Detection, SVM, Naive Bayes, SMOTE
A modified MixColumn-InversMixColumn in AES algorithm suitable for hardware implementation using FPGA device Prayitno, Ragiel Hadi; Latifah; Sudiro, Sunny Arief; Madenda, Sarifuddin; Harmanto, Suryadi
Communications in Science and Technology Vol 8 No 2 (2023)
Publisher : Komunitas Ilmuwan dan Profesional Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21924/cst.8.2.2023.1257

Abstract

This article described the Advanced Encryption Standard (AES) encryption and decryption process without using lookup tables in the MixColumns transformation and parallelizing the transformation process implemented in the Field Programmable Gate Array (FPGA) hardware. Parallelism of the hardware process conducted to the transformation of key schedule, addroundkey, subbyte and shiftrows (subshift) and mixcolumns in the first 5 rounds of the encryption process. The decryption process was parallelized in subshift transformations, both transformations were implemented at the same time. This research produced a modified AES encryption and decryption method and algorithm with the aim of minimizing the resources required for hardware implementation. The method in this article was applied to Xilinx ISE 14.7 software. The experimental results showed that the encryption process required 2,357 slice LUT's, 845 occupied slices and 26 IOB's, while the decryption process required 2,896 LUT's, 1,323 occupied slices and 26 IOB's resources. The encryption and decryption processes each took an average of 2.891 nanoseconds and 3.467 nanoseconds for every 128 bits of data. This approach leads us to obtain a component with minimum resources and enough computational speed.
Biometric System for Person Authentication Using Retinal Vascular Branching Pattern Diana Tri Susetianingtias; Sarifuddin Madenda; Rodiah; Rini Arianty
Jurnal Ilmu Komputer dan Informasi Vol. 16 No. 2 (2023): Jurnal Ilmu Komputer dan Informasi (Journal of Computer Science and Informatio
Publisher : Faculty of Computer Science - Universitas Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21609/jiki.v16i2.1156

Abstract

The person’s retina has its uniqueness that can be used as biometric recognition. The use of the retina as a marking feature in biometrics is more accurate in making calls, verification, and authentication. Retinal biometric characteristics are unique and difficult to manipulate, thus making the retinal biometric system one of the most reliable biometrics compared to other biometric characteristics. The retinal biometric system can be formed using extracted retinal vessels. The difficulty in extracting retinal vessels is a characteristic of retinal vessels. itself includes (central artery, central branch artery, central vein and central branch vein), the ratio of the thickness ratio between the different retinal arteries and veins (2:3), the location of the retinal artery and vein and the color. This complexity often results in errors in the retinal blood vessel extraction process, where not all blood vessel objects can be extracted properly which can reduce the accuracy of the retinal biometric system. This study will address the problem of extracting retinal vessels by proposing the use of an extraction method to produce truly unique retinal features to be included in the retinal biometric system by tracing all branches of the retinal vessels (consisting of: bifurcation, trifurcation and crossover). ). The accuracy results show that 99.81% of the images were correctly detected. The blood pattern is obtained by doing extraction which includes the preprocessing stage and is continued by doing the blood extraction stage. This pattern extraction result is used as a unique pattern to be included in the feature vector of the biometric system in identifying person based on the retina.
A New Chaos Function Developed through the Composition of the MS Map and the Circle Map Ichsani Mursidah; S Suryadi; Sarifuddin Madenda; Suryadi Harmanto
Proceeding International Conference on Mathematics and Learning Research 2023: Proceeding International Conference on Mathematics and Learning Research
Publisher : Universitas Muhammadiyah Surakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The rise of digital data theft makes researchers try to find better methods for digital data protection. Confidential digital data can be secured by encryption methods, one of which is the chaos function. We propose in this paper a new chaos function which is a composition of MS Map and Circle Map functions. This function has chaotic nature and is named the MS Circle Map. The sensitivity and randomness tests of the MS Circle map function are carried out using a bifurcation diagram, Lyapunov Exponent, and NIST. The analysis result of the bifurcation diagram shows that the MS Circle map has a good density at the value of r E (0,4). Besides that, the Lyapunov Exponent has a non-negative value at r E [0.4, 4], X0 = 0.9, r = 3.8, W = 0.5 λ = 2.1 which is the domain Xn E (0, 1) and parameter values r, E, and Ω, K are any real numbers. The results of the NIST randomness level test show that the MSC Map function all passed the randomness test of 16 NIST tests.
Penerapan Formula Mean dan Varian untuk Pengolahan Citra Menggunakan Perangkat FPGA Aqwam Rosadi Kardian; Sunny Arief Sudiro; Sarifuddin Madenda; Lingga Hermanto
Prosiding Seminar SeNTIK Vol. 1 No. 1 (2017): Prosiding SeNTIK 2017
Publisher : Lembaga Penelitian dan Pengabdian Kepada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Proses pengolahan citra membutuhkan rumus matematika yang diterapkan berdasarkan rumus pada data citra. Perangkat lunak mudah untuk melakukan hal itu dan mengakses data dari memori. Sebaliknya untuk implementasi perangkat keras dengan banyak kendala. Penelitian ini mengusulkan penerapan formula Mean dan Varian ke dalam Perangkat FGPA. Selain untuk perhitungan Optimal Mean dan Varian dengan hanya butuh satu komponen (dalam satu akumulator) dan satu pembagi menggunakan register geser kanan (Shift Right Register), untuk ukuran 8x8 citra diperlukan 64 clock putaran untuk menyelesaikan perhitungan mean dan varian
Algoritma Zigzag Scan Menggunakan Metode Pemetaan Untuk Mendukung Proses Kompresi Citra JPEG Robby Candra; Sarifuddin Madenda; Sunny Arief Sudiro
Prosiding Seminar SeNTIK Vol. 1 No. 1 (2017): Prosiding SeNTIK 2017
Publisher : Lembaga Penelitian dan Pengabdian Kepada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Kapasitas data multimedia terus bertambah, hal ini mengakibatkan jumlah data semakin besar sehingga membutuhkan jalur komunikasi yang besar. Salah satu cara untuk mengurangi ukuran data citra yaitu dengan cara memampatkan atau mengkompresi data citra, sehingga dapat meningkatkan kecepatan pengiriman data citra tanpa banyak mengurangi kualitas citra itu sendiri. Salah satu bagian dari algoritma JPEG yang dapat menentukan rasio kompresi dan kecepatan proses kompresi data adalah algoritma Zigzag Coding atau Zigzag Scan. Penelitian ini mengembangkan suatu teknik optimalisasi algoritma zigzag scan dengan metode pemetaan. Zigzag scan dengan metode pemetaan merupakan proses pengurutan data hasil DCT-terkuantisasi sesuai dengan urutan posisi yang sudah ditentukan secara zigzag. Proses pemetaan dilakukan antara urutan data masukan dan urutan posisi zigzag. Optimalisasi zigzag scan dengan metode pemetaan telah berhasil dikembangkan yaitu mampu meningkatkan waktu proses pengurutan koefisien-koefisien DCT-terkuantisasi yaitu 6 kali lebih cepat dari cara pengurutan berdasarkan pola zigzag yang ada saat ini, karena dengan menggunakan metode pemetaan data input langsung ditempatkan berdasarkan urutan posisi yang sudah ditentukan tanpa adanya perbandingan nilai dan tidak banyak proses perulangan
Analisa Metode Perbaikan Kualitas Citra Pada Citra Gigi Panoramik Rahayu Noveandini; Sarifuddin Madenda; Metty Mustikasari; Muhammad Subali
Prosiding Seminar SeNTIK Vol. 2 No. 1 (2018): Prosiding SeNTIK 2018
Publisher : Lembaga Penelitian dan Pengabdian Kepada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Citra gigi digunakan untuk identifikasi karena survivabilitas dan dapat memberikan banyak informasi tentang gigi seperti kerangka gigi, posisi gigi, bentuk gigi seperti adanya tambalan atau tidak dan sebagainya. Oleh sebab itu kandidat terbaik untuk identifikasi biometrik menggunakan fitur gigi [2]. Perbaikan kualitas citra menjadi kegiatan yang penting saat akan melakukan pengenalan suatu obyek citra. Tujuan utama perbaikan kualitas citra adalah mendapatkan hasil citra yang optimal dari citra aslinya sehingga memudahkan proses selanjutnya pada pengenalan suatu obyek. Beberapa metode perbaikan kualitas citra yang sering digunakan adalah histogram equalization (HE), adaptive histogram equalization (AHE) dan contras adaptive histogram equalization (CLAHE). Dalam penelitian ini, diusulkan perbaikan kualitas citra gigi panoramik dengan metode HE, AHE dan AHE dengan sharpening. Algoritma yang diusulkan akan membandingkan hasil ketiganya. Hasil dari penelitian ini menyimpulkan metode AHE dengan sharpening menghasilkan perbaikan kualitas citra yang paling optimal dilihat dari citra yang dihasilkan.