The goal of this project is to create a system that can identify a road sign's location in a picture and categorize it. We created functions to reduce dynamic range, modify image resolution, and convert RGB to grayscale for this purpose. Based on the outcomes of various clustering techniques, we chose to use the K-Means algorithm with five clusters. Additionally, 32x32 resolution photos have been selected as acceptable. A bounding box is created around the sign after selecting the appropriate structure to use. We created the code to operate on every input image because these procedures were initially performed on a single test image. Finally, by inheriting the feature extraction portion of VGG19, a transfer learning model has been created.
Eser Adı (dc.title) | Traffic Sign Detection |
Eser Sahibi (dc.contributor.author) | EGE ÖZBENDER |
Tez Danışmanı (dc.contributor.advisor) | Aytuğ ONAN |
Yayıncı (dc.publisher) | İzmir Katip Çelebi Üniversitesi Fen Bilimleri Enstitüsü |
Tür (dc.type) | Diğer |
Açıklama (dc.description) | Yüksek Lisans Bitirme Projesi |
Özet (dc.description.abstract) | The goal of this project is to create a system that can identify a road sign's location in a picture and categorize it. We created functions to reduce dynamic range, modify image resolution, and convert RGB to grayscale for this purpose. Based on the outcomes of various clustering techniques, we chose to use the K-Means algorithm with five clusters. Additionally, 32x32 resolution photos have been selected as acceptable. A bounding box is created around the sign after selecting the appropriate structure to use. We created the code to operate on every input image because these procedures were initially performed on a single test image. Finally, by inheriting the feature extraction portion of VGG19, a transfer learning model has been created. |
Kayıt Giriş Tarihi (dc.date.accessioned) | 2023-07-14 |
Açık Erişim Tarihi (dc.date.available) | 2023-07-14 |
Yayın Tarihi (dc.date.issued) | 2023 |
Tek Biçim Adres (dc.identifier.uri) | https://hdl.handle.net/11469/3470 |
Yayın Dili (dc.language.iso) | tr |
Konu Başlıkları (dc.subject) | Traffic Sign |
Konu Başlıkları (dc.subject) | Detection |