IMPROVING FINAL PREDICTION FROM INTERNAL ENSEMBLE LEARNING TRANSFORMER FOR OXFORD IIIT-PET DATASET - Dalam bentuk buku karya ilmiah
WESLI YEREMI VALENTINO RAMBI
Informasi Umum
Kode
25.05.670
Klasifikasi
000 - General Works
Jenis
Karya Ilmiah - Thesis (S2) - Reference
Subjek
Artificial Intelegence
Dilihat
27 kali
Informasi Lainnya
Abstraksi
Research in image classification has attracted considerable interest, par ticularly in Fine-Grained Visual Classification (FGVC), specializing in the com plex task of differentiating objects and subtle variations within species, such as those observed in different animal types. Various techniques have been developed to tackle these challenges, including feature coding, part-based approaches, and attention-based approaches. Within these approaches, the Vision Transformer (ViT) has shown remarkable success in image recognition tasks. The Internal Ensemble Learning Transformer (IELT) builds on ViT as its foundation, achieving impressive results. To enhance the effectiveness of IELT, we propose an innovative approach that focuses on refining its feature representation. This is achieved by incorporating a softmax activation function and a Radial Basis Function (RBF) layer to enhance the final prediction accuracy. Experimental findings demonstrate that our proposed method significantly boosts accuracy on fine-grained datasets, such as Oxford-IIIT Pet, Surpassing current cutting-edge methods. Keywords: Fine-Grained Visual Classification, Vision Transformer, Radial Basis Function
- TEI6A3 - SISTEM CERDAS
Koleksi & Sirkulasi
Tersedia 1 dari total 1 Koleksi
Anda harus log in untuk mengakses flippingbook
Pengarang
| Nama | WESLI YEREMI VALENTINO RAMBI |
| Jenis | Perorangan |
| Penyunting | Suryo Adhi Wibowo, Unang Sunarya |
| Penerjemah |
Penerbit
| Nama | Universitas Telkom, S2 Teknik Elektro |
| Kota | Bandung |
| Tahun | 2025 |
Sirkulasi
| Harga sewa | IDR 0,00 |
| Denda harian | IDR 0,00 |
| Jenis | Non-Sirkulasi |
