Islamic Geometric Recognition Using Support Vector Machine and SIFT Descriptor
Publish Year: 1395
نوع سند: مقاله کنفرانسی
زبان: English
View: 415
This Paper With 6 Page And PDF and WORD Format Ready To Download
- Certificate
- من نویسنده این مقاله هستم
استخراج به نرم افزارهای پژوهشی:
شناسه ملی سند علمی:
COMCONF03_201
تاریخ نمایه سازی: 6 اردیبهشت 1396
Abstract:
Pattern recognition and machine vision play a very crucial role in the present researches. In computer vision, shape classification is a fundamental problem with many applications like shape retrieval and object recognition. In this paper, a proposed system for recognize Islamic geometric patterns with emphasis on representation stages is introduced. In order to achieve this, a shape based classification method is used. So before classification, an efficient shape descriptor like SIFT has been utilized as feature extraction algorithm and then a recognition model is constructed base on a support vector machine machine. This work consists of three stages: image acquisition and preprocessing, feature extraction, and recognition. A set of different experiments on images of regular, translated, rotated and scaled Islamic geometric shape has been done and variety of results has been presented. As compared to other forms of classification model, the results of proposed recognition model is acceptable.
Keywords:
Key words: Islamic Geometric Patterns , SIFT descriptors , Support Vector Machine , The Recognition of Islamic Geometric Shapes
Authors
Narges Ahangar
Faculty of Electrical. Computer and IT Engineering, Islamic Azad University. Karaj Branch, Iran,
Azam Bastanfard
Faculty of Electrical. Computer and IT Engineering, Islamic Azad University. Karaj Branch, Iran,
مراجع و منابع این Paper:
لیست زیر مراجع و منابع استفاده شده در این Paper را نمایش می دهد. این مراجع به صورت کاملا ماشینی و بر اساس هوش مصنوعی استخراج شده اند و لذا ممکن است دارای اشکالاتی باشند که به مرور زمان دقت استخراج این محتوا افزایش می یابد. مراجعی که مقالات مربوط به آنها در سیویلیکا نمایه شده و پیدا شده اند، به خود Paper لینک شده اند :