Jihan Kamilah, . (2024) KLASIFIKASI CITRA REMPAH ADAS DAN JINTAN MENGGUNAKAN ALGORITMA CONVOLUTIONAL NEURAL NETWORK (CNN). Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.
![[img]](http://repository.upnvj.ac.id/style/images/fileicons/text.png) |
Text
ABSTRAK.pdf Download (31kB) |
|
Text
AWAL.pdf Download (574kB) |
|
Text
BAB 1.pdf Restricted to Repository UPNVJ Only Download (187kB) |
|
Text
BAB 2.pdf Restricted to Repository UPNVJ Only Download (430kB) |
|
Text
BAB 3.pdf Restricted to Repository UPNVJ Only Download (123kB) |
|
Text
BAB 4.pdf Restricted to Repository UPNVJ Only Download (2MB) |
|
Text
BAB 5.pdf Download (36kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (109kB) |
|
Text
RIWAYAT HIDUP.pdf Restricted to Repository UPNVJ Only Download (75kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository UPNVJ Only Download (1MB) |
|
Text
HASIL PLAGIARISME.pdf Restricted to Repository staff only Download (544kB) |
|
Text
ARTIKEL KI.pdf Restricted to Repository staff only Download (956kB) |
Abstract
Spices are aromatic plants used for seasoning, flavor enhancement, aroma, and food preservation with limited use. According to data from Badan Pusat Statistik in 2022 Indonesia exported 279306.9 tons of medicinal plants, aromatics, and spices to various countries. Indonesia has a long history associated with spices since the 15th century. Unfortunately, this isn't aligned with people's knowledge, especially generation Z, in identifying spices. There are spice sellers in traditional markets do not know certain spices. Thus, digital image processing techniques are needed to help identify the types of spices, especially seed spices. Method used in image classification is Convolutional Neural Network (CNN) with ResNet50 architecture using transfer learning and fine-tuning techniques. This research uses 320 spice images, consisting of 4 classes, which are anise, fennel, black cumin, and white cumin. Data splitting applied a ratio of 80:10:10 for train, val, and test data. In the training step, transfer learning technique model has a more efficient training time which is 3 hours 14 minutes, while the fine-tuning technique requires 4 hours 6 minutes for 100 epochs. Evaluation on test data showed both models achieving 100% accuracy, precision, recall, and f-1 score, with the transfer learning model having a loss of 0.0166 and the fine-tuning model a loss of 0.0028. However, in predicting new images, the fine-tuning model performed better with 91% correct predictions compared to 78% by the transfer learning model from 152 new image predictions.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Additional Information: | [No.Panggil: 2010511013] [Pembimbing 1: Nur Hafifah Matondang] [Pembimbing 2: Nurul Afifah Arifuddin] [Penguji 1 : Bayu Hananto] [Penguji 2: Indra Permana Solihin] |
| Uncontrolled Keywords: | Spices, Convolutional Neural Network, ResNet50, Transfer Learning, Fine-Tuning |
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software T Technology > T Technology (General) |
| Divisions: | Fakultas Ilmu Komputer > Program Studi Informatika (S1) |
| Depositing User: | JIHAN KAMILAH |
| Date Deposited: | 28 Aug 2024 08:00 |
| Last Modified: | 28 Aug 2024 08:00 |
| URI: | http://repository.upnvj.ac.id/id/eprint/31105 |
Actions (login required)
 |
View Item |