ارسل ملاحظاتك

ارسل ملاحظاتك لنا







Automatic Recognition System for Mechanical Hand Tools Using Convolutional Neural Networks

المصدر: مجلة بحوث جامعة تعز - سلسلة الآداب والعلوم الإنسانية والتطبيقية
الناشر: جامعة تعز
المؤلف الرئيسي: Alnowaini, Ghazi (Author)
مؤلفين آخرين: Al-Khateeb, Hesham (Co-Author) , Al-Basser, Amer (Co-Author)
المجلد/العدد: ع26
محكمة: نعم
الدولة: اليمن
التاريخ الميلادي: 2021
الشهر: مارس
الصفحات: 1 - 8
رقم MD: 1261408
نوع المحتوى: بحوث ومقالات
اللغة: الإنجليزية
قواعد المعلومات: EduSearch, AraBase, HumanIndex
مواضيع:
كلمات المؤلف المفتاحية:
Mechanical Tools | Hand Tools | Deep Learning | CNN | Resnet
رابط المحتوى:
صورة الغلاف QR قانون
حفظ في:
LEADER 02527nam a2200265 4500
001 2013459
041 |a eng 
044 |b اليمن 
100 |9 672236  |a Alnowaini, Ghazi   |e Author 
245 |a Automatic Recognition System for Mechanical Hand Tools Using Convolutional Neural Networks 
260 |b جامعة تعز  |c 2021  |g مارس 
300 |a 1 - 8 
336 |a بحوث ومقالات  |b Article 
520 |b The identification of high-precision mechanical tools is an important problem faces mechanical engineers. Indeed, hundreds of different instruments are typically used for one task. In several cases and multiple workshop environments each tool will be used. Mechanical engineering is a practical field that need different mechanical tools during workshop work. These tools such as Wrench, hammer, toolbox, Gasoline Can, and pebble have different size and style. During work time of the mechanical engineers, they need these tools frequently and the identification process of these tools is a difficult task for automated system. In this paper, an automated recognition system for mechanical tools using convolution neural networks. ACNN-based model is discussed using four versions of residual network classifiers; ResNet-18, ResNet-34, ResNet-50 and ResNet-152. This model can be integrated with a robot to give it the ability to recognize the specific mechanical tool and deliver it to the mechanical engineer. The feasibility of this method is illustrated in the achieved results. The obtained results are very promising to be used in practical use. In term of testing accuracy, the results achieved 84%, 85%, 86% and 87% for ResNet18, ResNet-34, ResNet-50 and ResNet-152 respectively. 
653 |a الهندسة الميكانيكية  |a المهندسين الميكانيكيين  |a الشبكات العصبية  |a التعلم الآلي 
692 |b Mechanical Tools  |b Hand Tools  |b Deep Learning  |b CNN  |b Resnet 
700 |9 672239  |a Al-Khateeb, Hesham   |e Co-Author 
700 |9 672241  |a Al-Basser, Amer  |e Co-Author 
773 |4 الادب  |4 العلوم الاجتماعية ، متعددة التخصصات  |6 Literature  |6 Social Sciences, Interdisciplinary  |c 002  |e University of Taiz Research Journal - Arts and Humanities  |l 026  |m ع26  |o 0931  |s مجلة بحوث جامعة تعز - سلسلة الآداب والعلوم الإنسانية والتطبيقية  |v 000 
856 |u 0931-000-026-002.pdf 
930 |d n  |p y  |q n 
995 |a EduSearch 
995 |a AraBase 
995 |a HumanIndex 
999 |c 1261408  |d 1261408 

عناصر مشابهة