IPB University Logo

SCIENTIFIC REPOSITORY

IPB University Scientific Repository collects, disseminates, and provides persistent and reliable access to the research and scholarship of faculty, staff, and students at IPB University

AI Repository
 
Building and Categories


      View Item 
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Computer Science
      • View Item
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Computer Science
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Lokalisasi Kepala Sapi Otomatis Menggunakan Arsitektur YOLO Berbasis Transfer Learning

      Thumbnail
      View/Open
      Cover (763.9Kb)
      Fulltext (1.980Mb)
      Lampiran (817.5Kb)
      Date
      2026
      Author
      DJAFAR, JUSTIN KRISTALDI
      Hasibuan, Lailan Sahrina
      Metadata
      Show full item record
      Abstract
      Identifikasi individu sapi merupakan elemen krusial dalam manajemen peternakan modern. Namun, pengembangan model identifikasi sering terhambat oleh kebutuhan lokalisasi data manual yang memakan waktu dan tenaga. Penelitian ini bertujuan membangun sistem otomatisasi lokalisasi citra kepala sapi menggunakan arsitektur YOLOv11 yang dilatih dengan teknik transfer learning dari bobot pretrained COCO. Data berupa rekaman video dari 24 individu sapi diperoleh dari Peternakan Koperasi Produsen Karya Nugraha Jaya, Kuningan, yang selanjutnya diekstraksi menjadi 1.808 citra melalui teknik sampling temporal sebesar 1 citra per detik menggunakan pustaka OpenCV. Dataset dibagi secara cow-disjoint dengan proporsi 70:15:15 untuk training, validation, dan testing, sehingga estimasi kinerja mencerminkan kondisi deployment pada individu yang belum dikenal sistem. Untuk memperoleh perbandingan yang robust, empat varian arsitektur YOLOv11 (n, s, m, l) masing-masing dilatih selama 80 epoch dengan ukuran citra 640 piksel dan batch size 8. Evaluasi kinerja model dilakukan menggunakan metrik Success Rate pada ambang IoU 0,75 (SR@0.75) untuk mengukur proporsi anotasi bounding box yang memenuhi tingkat kesesuaian tinggi terhadap ground truth. Hasil evaluasi digunakan untuk menganalisis kemampuan masing-masing varian YOLOv11 dalam menghasilkan anotasi otomatis yang akurat, sehingga dapat menjadi dasar dalam pengembangan sistem lokalisasi otomatis untuk mempercepat pembangunan dataset identifikasi ternak.
       
      Individual cattle identification is a crucial component of modern livestock management. However, the development of identification models is often hindered by the time-consuming and labor-intensive process of manual data annotation. This study aims to develop an automated cattle head annotation system using the YOLOv11 architecture trained through transfer learning from pretrained COCO weights. Video recordings of 24 individual cattle were collected from Koperasi Produsen Karya Nugraha Jaya Farm, Kuningan, and subsequently converted into 1,808 images through temporal sampling at an interval of one frame per second using the OpenCV library. The dataset was partitioned using a cow-disjoint strategy with a 70:15:15 ratio for training, validation, and testing, ensuring that the performance evaluation reflects deployment scenarios involving previously unseen individuals. To obtain a robust comparison, four YOLOv11 variants (n, s, m, and l) were each trained for 80 epochs using an input image size of 640 pixels and a batch size of 8. Model performance was evaluated using the Success Rate at an Intersection over Union threshold of 0.75 (SR@0.75), which measures the proportion of predicted bounding boxes that achieve a high level of agreement with the ground-truth annotations. The evaluation results were analyzed to compare the localization performance of each YOLOv11 variant in generating accurate automatic annotations, providing a foundation for the development of automated localization systems that can accelerate the creation of livestock identification datasets.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/178953
      Collections
      • UF - Computer Science [194]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository