Claim Missing Document
Check
Articles

Found 27 Documents
Search

Penggunaan Metode Haar Cascade Classifier dan LBPH Untuk Pengenalan Wajah Secara Realtime Ramadini, Febrin Ludia; Haryatmi, Emy
InfoTekJar : Jurnal Nasional Informatika dan Teknologi Jaringan Vol 6, No 2 (2022): InfoTekJar Maret
Publisher : Universitas Islam Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30743/infotekjar.v6i2.4714

Abstract

Pengenalan wajah manusia menjadi sebuah topik penelitian biometric yang cukup banyak diminatai karena pada wajah manusia terdapat banyak informasi terutama mengenai identitas seseorang. Setiap orang memiliki bentuk wajah yang berbeda yang dapat dilihat dari mata, hidung, telinga dan juga mulut. Pada penelitian ini penulis menggabungkan dua metode haar cascade classifoer dan LBPH untuk membuat sistemm yang dapat mengenali wajah seseorang. Metode haar casecade classifier digunakan untuk mendeteksi adanya wajah manusia sedangkan metode LBPH digunakan untuk mengenali wajah seseorang. Pada sistm ini terdapat beberapa proses untuk dapat mengenali wajah seseorang, yaitu: proses deteksi wajah, proses pengambilan dataset, proses pelatihan wajah dan proses pengenalan wajah. Proses pengambilan dataset dilakukan secara otomatis, saat sistem sudah mendeteksi adanya wajah manusia dan diambil sebanyak 40 foto untuk setiap satu wajah user. Sistem akan mencocokkan wajah yang terdeteksi dengan indentitas wajah yang telah dimasukkan ke dalam dataset. Selanjutnya sisstem akan mengenali wajah yang dideteksi dan menampikan nama sesuai dengan nomer user ID yang terdapat di dataset. Tampilan pengenalan wajah mnggunakan sistem realtime dimana nama yang ditampilkan sesuai dengan orang yang tepat berdiri didepan kamera laptop pada saat itu. Keberhasilan sistem ini sebesar 88,42%.
Design and Implementation of FTTB Network Transmission in High-Rise Buildings Using GPON Ahmad Yuhani; Emy Haryatmi; Deni Haryadi; Yunus Bakhtiar Arafat
International Journal of Engineering Continuity Vol. 4 No. 1 (2025): ijec
Publisher : Sultan Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58291/ijec.v4i1.375

Abstract

The advancement of digital communication technology has significantly increased the demand for reliable network infrastructure, particularly in high-rise buildings such as hotels and resorts. This study aims to design and evaluate a Fiber To The Building (FTTB) network system based on Gigabit Passive Optical Network (GPON) technology at The Anvaya Beach Resort Bali. The system is designed to distribute integrated communication services including data voice, and video, through a zoning approach based on the functional layout of the building. The Waterfall method is employed in the system development, encompassing the stages of requirement analysis, topology design, field implementation, and optical performance testing. The findings indicate that most zones have optical attenuation values within the standard range (15–28 dB), and the received signal power remains within the acceptable threshold (-28 dBm). However, several areas exhibit suboptimal signal performance, particularly those with long distribution paths and a high number of optical splitters. Zone C1.1 demonstrates the best performance, with stable attenuation levels and signal strength within standards, without requiring additional active devices. The study concludes that a GPON-based FTTB system can efficiently and flexibly meet the data communication needs of high-rise buildings and other complex building environments with similar systems.
Rancang Bangun Sistem Pengering Sepatu Dengan Menggunakan Aplikasi Telegram Berbasis ESP32 Bintang Amalia Putri; Emy Haryatmi; Tri Agus Riyadi; Ary Bima Kurniawan
Jurnal Ilmiah Informatika Komputer Vol. 31 No. 1 (2026)
Publisher : Universitas Gunadarma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35760/ik.2026.v31i1.632

Abstract

Sepatu merupakan salah satu alas kaki yang banyak digunakan untuk beraktivitas diluar ruangan. Ketika musim hujan tiba, terkadang sepatu menjadi basah dan tidak dapat digunakan kembali di hari berikutnya. Kondisi sepatu yang basah juga dapat disebabkan oleh aktivitas luar ruangan. Sepatu yang basah dapat menyebabkan bau tidak sedap, pertumbuhan jamur, serta kerusakan material sepatu. Proses pengeringan secara konvensional sering kali kurang efektif karena bergantung pada cuaca dan memerlukan waktu yang lama. Penelitian ini bertujuan untuk merancang dan membuat alat pengering sepatu otomatis dengan menggunakan aplikasi telegram untuk mempercepat proses pengeringan dan pemantauan kondisi sepatu secara real-time. Pada penelitian ini, jenis sepatu yang digunakan adalah sepatu berbahan kanvas, synthetic leather dan mesh untuk sepatu dewasa dan sistem akan mati secara otomatis berdasarkan timer yang sudah ditentukan sebelumnya dan disesuaikan dengan nilai kelembabannya atau dari aplikasi Telegram. Metode yang digunakan meliputi perancangan sistem berbasis mikrokontroler ESP32 yang terhubung dengan sensor DHT11 untuk mengukur suhu dan kelembaban, serta modul relay untuk mengendalikan kipas, lampu pijar, dan perangkat shoe dryer. Sistem ini dilengkapi dengan tampilan LCD dan terhubung dengan aplikasi Telegram untuk kontrol jarak jauh serta notifikasi otomatis. Hasil pengujian menunjukkan bahwa bahan sepatu seperti synthetic leather dan mesh dapat mengering sempurna dengan nilai kelembaban adalah 0% dalam waktu 3 jam, sedangkan sepatu dengan bahan kanvas membutuhkan waktu lebih lama yaitu 3 jam 23 menit dengan kelembaban sekitar 7–9%. Sistem ini dapat mengeringkan sepatu terutama bagian dalamnya dengan sempurna.
Implementation of Intrusion Detection System Using Snort and Log Visualization Using ELK Stack Fatih Dien Robbani; Emy Haryatmi; Tri Agus Riyadi; Riza Adrianti Supono; Ary Bima Kurniawan; Rosdiana Rosdiana
International Journal of Engineering, Science and Information Technology Vol 5, No 3 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i3.901

Abstract

Cyber threats like malware, ransomware, and DDoS attacks demand fast and integrated detection systems. Traditional network monitoring tools often struggle to identify complex real-time attack patterns. This study evaluates the integration of Snort, an Intrusion Detection System (IDS), with the ELK Stack (Elasticsearch, Logstash, Kibana) to detect and visualize cyberattacks effectively. The system was tested against three attack scenarios: a Windows ping flood, port scanning using Zenmap, and SSH brute force attacks via Nmap Scripting Engine (NSE). Wireshark was employed as a supporting tool to monitor raw network traffic. The results indicate that Snort detected all simulated attacks in real time, and the ELK Stack efficiently processed and visualized the alert data. However, limitations in Kibana's dashboard refresh rate slightly hindered real-time monitoring capabilities. Overall, the integration of Snort and the ELK Stack proves effective for network threat detection and analysis, with room for future improvements in visualization performance and automated response mechanisms.
Herbal Plant Leaves Classification Using Convolutional Neural Network Models: A Literature Review Alfharizky Fauzi; Emy Haryatmi; Tri Agus Riyadi; Murniyati Murniyati
International Journal of Engineering, Science and Information Technology Vol 5, No 1 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i1.723

Abstract

Plants are essential to human beings because plants are considered most as foods. Plants can be used for food ingredients, medical purposes, and industrial applications. People inspect plants using traditional methods, such as using the naked eye, which can be time-consuming and expensive. Therefore, the effectiveness and high quality of automated crop identification classification systems are needed for adequate crop protection. This study aims to identify and classify nine plant species using different datasets, focusing on transfer learning from models trained on plant leaf datasets. Most research has shown that increasing the dataset size would significantly improve classification accuracy. The accuracy of the first test using the modified N1 classification model was 99.45%. In the second experiment, the accuracy of the N2 model was 99.65%. The accuracy of the N3 model, despite being slightly less accurate than AlexNet, was 99.55%, and it performed better, while the accuracy of AlexNet was 99.73%. Compared to the AlexNet model, the proposed model performed better and required less training time. The N1 model reduced the training time by 34.58%, the N2 model by 18.25%, and the N3 model by 20.23%. The N1 and N3 resulted in the same size, namely 14.8MB, and the compactness was 92.67%. The size of the N2 model was 29.7MB, and the compactness was 85.29% compactness. The proposed models provide more accuracy and efficiency in classifying plant leaves and can be used as a standalone mobile application that benefits farmers.
Classification of Tomato Ripeness Based on Convolutional Neural Network Methods Nur Azizah Ayunda; Emy Haryatmi; Tri Agus Riyadi
Journal of Information System and Informatics Vol 5 No 4 (2023): Journal of Information Systems and Informatics
Publisher : Asosiasi Doktor Sistem Informasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51519/journalisi.v5i4.613

Abstract

Sorting system for tomato is one of the important things to deploy to achieve better quality of tomato. Nowadays, many sorting system is done manually and this could spend a lot of time and become inefficient. One method can be implemented in the sorting system by using Convolutional Neural Network (CNN) method to classify the ripeness of tomatoes. The objective of this research is to classify the ripeness of tomatoes based on the color of tomatoes. There are three categories of color level such as green for raw tomato, turning for half-ripe tomato and red for ripe tomato. Research methodology of this research is data collection, data pre-processing and image maintenance, CNN model, and training data. The image used in this research are 1148 images. These images were taken manually using smartphone camera in outdoor environment. These images were used to build CNN model. The results of this research show that by testing 10 images of tomatoes achieved raw tomatoes close to 90%, ripe tomatoes close to 90% and half-ripe tomatoes close to 80%. Based on the results, CNN can be used as a good alternative in image classification tasks.
A Comprehensive Literature Review on Skin Disease Identification Through Deep Learning-Based Bi-Directional Image-Text Matching Techniques Alfharizky Fauzi; Adang Suhendra; Aries Muslim; Emy Haryatmi
International Journal of Engineering, Science and Information Technology Vol 6, No 4 (2026)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v6i4.1689

Abstract

Over the past few years, matching visual and textual data commonly known as image text matching has become increasingly important due to the rising demand for multimodal data understanding. ITM involves retrieving an image that matches a textual description, or vice versa, by evaluating the semantic similarity between visual and textual inputs. This task becomes particularly critical in medical applications such as skin disease identification, where accurate matching between clinical images and descriptive diagnostic text can aid in early detection and decision-making. One of the core challenges in this domain lies in bridging the semantic gap between visual features of skin lesions and medical terminology used in clinical notes or diagnostic reports. Despite various deep learning approaches developed for general ITM tasks, limited research has focused on domain-specific applications like dermatology. This paper provides a focused review of deep learning-based ITM frameworks, emphasizing their relevance to skin disease analysis. We explore model components such as image and text encoders, common embedding strategies (e.g., global, local, hybrid), and learning paradigms including adversarial, ranking, and pre-trained vision-language models. Additionally, benchmark datasets are discussed, including synthetic and real-world skin image–text pairs. The paper concludes by identifying key challenges such as dataset scarcity, domain adaptation, and explainability highlighting opportunities for advancing ITM in dermatological diagnostics.