@prefix skos: . @prefix geo: . @prefix rdf: . @prefix cc: . @prefix bibo: . @prefix void: . @prefix event: . @prefix foaf: . @prefix eprel: . @prefix rdfs: . @prefix dc: . @prefix dct: . @prefix epid: . @prefix xsd: . @prefix ep: . @prefix owl: . <> foaf:primaryTopic ; rdfs:comment "The repository administrator has not yet configured an RDF license."^^xsd:string . dc:format "text/html"; dc:title "HTML Summary of #11061 \n\nPEMODELAN KLASIFIKASI PENYAKIT PADA DAUN ANGGUR MENGGUNAKAN \nCONVOLUTIONAL NEURAL NETWORK (CNN)\n\n"; foaf:primaryTopic . rdf:_1 . , ; bibo:abstract "PEMODELAN KLASIFIKASI PENYAKIT PADA DAUN ANGGUR MENGGUNAKAN\r\nCONVOLUTIONAL NEURAL NETWORK (CNN)\r\nNovia Permata Sari1\r\n; Sufiatul Maryana2 ; Adriana Sari Aryani3\r\n.\r\nProgram Studi Ilmu Komputer1,2,3\r\nUniversitas Pakuan, Bogor, Indonesia1,2,3\r\nwww.unpak.ac.id1,2,3\r\nNopiapermatasar518@gmail.com1\r\n, sufiatul.maryana@unpak.ac.id2\r\n, adriana.aryani@unpak.ac.id3\r\n(*) Corresponding Author\r\nThe creation is distributed under the Creative Commons Attribution-NonCommercial 4.0 International Lincense.\r\nAbstract- Grapevine diseases often pose a\r\nmajor challenge, leading to reduced crop\r\nquality and yield if not detected early. Manual\r\ndisease identification typically requires\r\nspecialized expertise, high precision, and\r\nconsiderable time. This study aims to develop\r\nan automated deep learning-based\r\nclassification model for grapevine leaf\r\ndiseases using the Convolutional Neural\r\nNetwork (CNN) method with a DenseNet121\r\nbase architecture. The dataset consists of\r\n1,600 digital images of green grapevine\r\nleaves (*Vitis vinifera* L. var. Thompson\r\nSeedless) collected from plantations in\r\nSukaraja District, Bogor Regency. These\r\nimages are evenly distributed across four\r\ncondition classes: Healthy, Black Measles\r\n(Esca), Leaf Blight (Isariopsis Leaf Spot),\r\nand Black Rot. Model testing was conducted\r\nusing two data split scenarios: 80:10:10 and\r\n70:15:15. The results demonstrate that the\r\nDenseNet121 architecture effectively\r\nextracts visual image characteristics. The best\r\nperformance was achieved with the 70:15:15\r\ndata split, yielding a validation accuracy of\r\n91.25%, whereas the 80:10:10 split resulted\r\nin an accuracy of 88.75%. The Healthy and\r\nLeaf Blight classes were identified with\r\n100% accuracy. This classification model\r\nwas integrated into a web application built on\r\nthe Flask framework as a practical\r\nimplementation to support smart a validation\r\naccuracy of 91.25%, whereas the 80:10:10\r\nsplit resulted in an accuracy of 88.75%. The\r\nHealthy and Leaf Blight classes were\r\nidentified with 100% accuracy. This\r\nclassification model was integrated into a\r\nweb application built on the Flask framework\r\nas a practical implementation to support\r\nsmart agriculture systems.\r\nKeywords: Convolutional Neural Network\r\n(CNN), Deep Learning, Image Classification,\r\nGrape Leaf Disease, DenseNet121.\r\nAbstrak- Penyakit pada tanaman anggur\r\nsering menjadi kendala utama yang\r\nmenyebabkan penurunan kualitas dan\r\nkuantitas hasil panen apabila tidak dideteksi\r\nsejak dini. Identifikasi penyakit secara manual\r\numumnya membutuhkan keahlian khusus,\r\nketelitian tinggi, dan waktu yang relatif lama.\r\nPenelitian ini bertujuan untuk membangun\r\nmodel klasifikasi otomatis penyakit pada daun\r\nanggur berbasis Deep Learning menggunakan\r\nmetode Convolutional Neural Network\r\n(CNN) dengan arsitektur dasar DenseNet121.\r\nDataset yang digunakan berupa 1.600 citra\r\ndigital daun anggur hijau (Vitis vinifera L.\r\nvar. Thompson Seedless) yang dikumpulkan\r\ndari perkebunan di Kecamatan Sukaraja,\r\nKabupaten Bogor. Citra tersebut terbagi rata\r\nke dalam empat kelas kondisi: Healthy\r\n(Sehat), Black Measles (Esca), Leaf Blight\r\n(Isariopsis Leaf Spot), dan Black Rot.\r\nPengujian model dilakukan menggunakan dua\r\nskenario pembagian proporsi data, yaitu\r\n80:10:10 dan 70:15:15. Hasil penelitian\r\nmenunjukkan bahwa arsitektur DenseNet121\r\nmampu mengekstraksi karakteristik visual\r\ncitra secara optimal. Performa terbaik dicapai\r\npada skenario proporsi data 70:15:15 dengan\r\ntingkat akurasi validasi sebesar 91,25% ,\r\nsedangkan proporsi 80:10:10 menghasilkan\r\nakurasi sebesar 88,75%. Kelas Healthy dan\r\nLeaf Blight berhasil diidentifikasi secara\r\nsempurna dengan akurasi 100%. Model\r\nklasifikasi ini diintegrasikan ke dalam aplikasi\r\nweb berbasis framework Flask sebagai\r\nimplementasi praktis pendukung sistem\r\npertanian cerdas (smart agriculture)."^^xsd:string; bibo:authorList ; bibo:status ; dct:creator , , , ; dct:date "2026-06-21"; dct:isPartOf ; dct:issuer , ; dct:subject ; dct:title "PEMODELAN KLASIFIKASI PENYAKIT PADA DAUN ANGGUR MENGGUNAKAN\r\nCONVOLUTIONAL NEURAL NETWORK (CNN)"^^xsd:string; rdf:type bibo:Article, bibo:Thesis, ep:EPrint, ep:ThesisEPrint; rdfs:seeAlso . rdf:type skos:Concept; skos:prefLabel "Ilmu Komputer"@en . foaf:name "Fakultas Matematika dan Ilmu Pnegetahuan Alam"^^xsd:string; rdf:type foaf:Organization . dct:hasPart ; foaf:name "Universitas Pakuan"^^xsd:string; rdf:type foaf:Organization . dct:isPartOf ; foaf:name "Fakultas Matematika dan Pengetahuan Alam, Universitas Pakuan"^^xsd:string; rdf:type foaf:Organization . foaf:name "Program Studi Ilmu Komputer"^^xsd:string; rdf:type foaf:Organization . foaf:familyName "Aryani"^^xsd:string; foaf:givenName "Adriana Sari"^^xsd:string; foaf:name "Adriana Sari Aryani"^^xsd:string; rdf:type foaf:Person . foaf:familyName "Sari"^^xsd:string; foaf:givenName "Novia Permata"^^xsd:string; foaf:name "Novia Permata Sari"^^xsd:string; rdf:type foaf:Person . foaf:familyName "Maryana"^^xsd:string; foaf:givenName "Sufiatul"^^xsd:string; foaf:name "Sufiatul Maryana"^^xsd:string; rdf:type foaf:Person .