Muhammad Zakaria Saputra, - (2024) KLASIFIKASI TANAMAN HIAS MENGGUNAKAN INCEPTION V3 PADA FITUR PENCARIAN GAMBAR DI APLIKASI HIAZEE. S1 thesis, Universitas Pendidikan Indonesia.
Abstract
Pengembangan usaha tanaman hias memiliki peluang besar di Indonesia dengan permintaan domestik yang tumbuh 21,8% per tahun dan nilai ekspor Rp1,3 triliun pada 2022. Minimnya pengetahuan masyarakat dalam mengenali jenis tanaman hias membuat pencarian teks di e-commerce kurang efektif, sehingga pencarian berbasis gambar menjadi solusi menarik. Penelitian ini mengembangkan model klasifikasi tanaman hias menggunakan transfer learning dengan arsitektur Inception V3 dan optimasi threshold untuk memaksimalkan F1-score, serta mengintegrasikan model ini ke fitur pencarian gambar di aplikasi Hiazee. Data tanaman hias diperoleh dari dataset Goletplant di Roboflow, melalui beberapa tahap pemrosesan, termasuk augmentasi, untuk meningkatkan kualitas model. Model Inception V3 yang telah dilatih pada dataset ImageNet digunakan sebagai model dasar dan diadaptasi dengan dataset spesifik tanaman hias melalui metode transfer learning. Model ini berhasil mengenali 36 jenis tanaman hias dengan akurasi 87,7% dan diintegrasikan ke fitur pencarian gambar. Hasil eksperimen menunjukkan bahwa kondisi lingkungan mempengaruhi akurasi model. The development of the ornamental plant business has significant potential in Indonesia, with domestic demand growing at 21.8% per year and export values reaching IDR 1.3 trillion in 2022. The lack of public knowledge in identifying various types of ornamental plants makes text-based searches on e-commerce platforms less effective, making image-based search a compelling solution. This research developed a classification model for ornamental plants using transfer learning with the Inception V3 architecture and threshold optimization to maximize the F1-score, integrating the model into the image search feature of the Hiazee application. Ornamental plant data was obtained from the Goletplant dataset on Roboflow and underwent several processing stages, including augmentation, to enhance model quality. The Inception V3 model, pre-trained on the ImageNet dataset, was used as the base model and adapted to the specific ornamental plant dataset using the transfer learning method. The model successfully recognized 36 types of ornamental plants with an accuracy of 87.7% and was integrated into the image search feature. Experimental results showed that environmental conditions affect the model’s accuracy.
![[img]](http://repository.upi.edu/style/images/fileicons/text.png) |
Text
S_KOM_2007993_Title.pdf Download (541kB) |
|
Text
S_KOM_2007993_Chapter1.pdf Download (294kB) |
|
Text
S_KOM_2007993_Chapter2.pdf Restricted to Staf Perpustakaan Download (2MB) |
|
Text
S_KOM_2007993_Chapter3.pdf Download (280kB) |
|
Text
S_KOM_2007993_Chapter4.pdf Restricted to Staf Perpustakaan Download (4MB) |
|
Text
S_KOM_2007993_Chapter5.pdf Download (227kB) |
|
Text
S_KOM_2007993_Appendix.pdf Restricted to Staf Perpustakaan Download (1MB) |
| Item Type: | Thesis (S1) |
|---|---|
| Additional Information: | https://scholar.google.com/citations?user=LIXMo1AAAAAJ&hl=en ID SINTA Dosen Pembimbing: Lala Septem Riza: 5975668 Muhammad Nursalman: 6143456 |
| Uncontrolled Keywords: | Klasifikasi Multilabel, Transfer Learning, Inception V3, Optimasi Thresholding, F1-score, E-commerce, Tanaman Hias. Multilabel Classification, Transfer Learning, Inception V3, Threshold Optimization, F1-score, E-commerce, Ornamental Plants. |
| Subjects: | L Education > L Education (General) Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software Q Science > QK Botany |
| Divisions: | Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Ilmu Komputer |
| Depositing User: | Muhammad Zakaria Saputra |
| Date Deposited: | 06 Sep 2024 01:28 |
| Last Modified: | 06 Sep 2024 01:28 |
| URI: | http://repository.upi.edu/id/eprint/123140 |
Actions (login required)
 |
View Item |