Writer Adaptation For End-To-EndArabic Online HandwritingRecognition / Fakhraddin Ahmed Ali Al-wajih ; Sherif Abdou,Eman Badr
Material type:
- text
- Unmediated
- volume
- ﺗوﻓﯾق ﻧظم اﻟﺗﻌرف اﻟﺷﺎﻣﻠﺔ ﻋﻠﻰ اﻟﻛﺗﺎﺑﺔ اﻟﻌرﺑﯾﺔ ﻣﻊ ﺧط اﻟﻛﺎﺗب
- 001.5561
Item type | Current library | Home library | Call number | Status | Barcode | |
---|---|---|---|---|---|---|
![]() |
قاعة الرسائل الجامعية - الدور الاول | المكتبة المركزبة الجديدة - جامعة القاهرة | Cai01.20.01.Ph.D.2022.Fa.W (Browse shelf(Opens below)) | Not for loan | 01010110086904000 |
Browsing المكتبة المركزبة الجديدة - جامعة القاهرة shelves Close shelf browser (Hides shelf browser)
No cover image available | No cover image available | No cover image available | No cover image available | No cover image available | No cover image available | No cover image available | ||
Cai01.20.01.Ph.D.2021.So.P Proposed system for image forgery detection / | Cai01.20.01.Ph.D.2021.So.P Proposed system for image forgery detection / | Cai01.20.01.Ph.D.2021.So.P Proposed system for image forgery detection / | Cai01.20.01.Ph.D.2022.Fa.W Writer Adaptation For End-To-EndArabic Online HandwritingRecognition / | Cai01.20.01.Ph.D.2022.Ka.P. Psychological human traits detection using deeppattern classification approaches / | Cai01.20.01.Ph.D.2023.Gh.A. Audio-visual matching assisted in sources speech separation / | Cai01.20.02.M.Sc.2003.Ah.M A model to solve multi - cycles machine time scheduling problems / |
Thesis (Ph.D)-Cairo University,Faculty of Computers and Artificial Intelligence,Department of Information Technology.
Bibliography: p. 110-112.
اﻟﻜﺘﺎﺑﺔ اﻟﯿﺪوﯾﺔ ھﻲ وﺳﯿﻠﺔ اﺗﺼﺎل ﻣﺮﻏﻮﺑﺔ ﺑﯿﻦ اﻟﺒﺸﺮ وأﺟﮭﺰة اﻟﻜﻤﺒﯿﻮﺗﺮ. ﻓﻲ اﻵوﻧﺔ اﻷﺧﯿﺮة، اﻛﺘﺴﺒﺖ أﻧﻈﻤﺔ اﻟﺘﻌﺮف ﻋﻠﻰ ﺧﻂ اﻟﯿﺪ اھﺘﻤﺎﻣﺎً أﻛﺒﺮ ﺑﺴﺒﺐ اﻟﺘﻘﺪم ﻓﻲ ﺗﻜﻨﻮﻟﻮﺟﯿﺎ اﻷﺟﮭﺰة اﻟﻤﺤﻤﻮﻟﺔ وأﯾﻀﺎً ﻟﺘﻘﺪم اﻟﺘﻌﻠﻢ اﻟﻌﻤﯿﻖ اﻟﻤﺴﺘﺨﺪم ﻓﻲ ھﺬه اﻷﻧﻈﻤﺔ. ﯾﺴﺘﻔﯿﺪ ﺗﻜﯿﯿﻒ اﻟﻜﺎﺗﺐ ﻣﻦ ﻧﻈﺎم اﻟﺘﻌﺮف ﻋﻠﻰ اﻟﻜﺘﺎﺑﺔ اﻟﯿﺪوﯾﺔ ﻟﻠﻜﺎﺗﺐ اﻟﻤﺴﺘﻘﻞ وﯾﺤﻮﻟﮭﺎ إﻟﻰ ﻧﻈﺎم ﯾﻌﺘﻤﺪ ھﺬه اﻟﺮﺳﺎﻟﺔ ﻋﺪداً ﻣﻦ اﻟﻨﻤﺎذج اﻟﺠﺪﯾﺪة اﻟﺸﺎﻣﻠﺔ ﺗﻘﺘﺮح ﻟﺘﺤﺴﯿﻦ أداء اﻟﺘﻌﺮف ﻋﻠﻰ ﺧﻂ اﻟﯿﺪ ﺑﺎﻟﻠﻐﺔ اﻟﻌﺮﺑﯿﺔ وﺗﻘﺘﺮح ﺗﻜﯿﯿﻔﺎً ﺟﺪﯾﺪاً ﻟﻠﻜﺎﺗﺐ ﻟﻠﻨﻤﺎذج اﻟﺸﺎﻣﻠﺔ ﻣﻦ ﺧﻼل اﻻﺳﺘﻔﺎدة ﻣﻦ ﻗﻮة اﻟﺸﺒﻜﺎت اﻟﻌﺼﺒﯿﺔ اﻟﻌﻤﯿﻘﺔ. أوﻻً، ﻧﺤﻦ ﻧﻘﺘﺮح ﻧﻈﺎم ﺗﻌﺮف ﻋﻠﻰ ﺧﻂ اﻟﯿﺪ ﺷﺎﻣﻞ ﻟﻠﻜﺘﺎﺑﺔ اﻟﻌﺮﺑﯿﺔ ﯾﻌﺘﻤﺪ ﻋﻠﻰ اﻟﺬاﻛﺮة ﺛﻨﺎﺋﯿﺔ اﻻﺗﺠﺎه طﻮﯾﻠﺔ المدى ﺗﻢ دﻣﺞ ﻧﻤﻮذج اﻟﻠﻐﺔ اﻹﺣﺼﺎﺋﯿﺔ N-gram ﻣﻊ ﻧﻤﻮذج اﻟﺸﺒﻜﺔ اﻟﻌﺼﺒﯿﺔ ﻹﺿﺎﻓﺔ اﻟﻤﻌﺮﻓﺔ اﻟﻠﻐﻮﯾﺔ وﺗﻌﺰﯾﺰ اﻷداء اﻟﻌﺎم ﻟﻠﻨﻈﺎم اﻟﻤﻘﺘﺮح. اﻟﻨﻈﺎم اﻟﻤﻘﺘﺮح ﻗﺎدر ﻋﻠﻰ أداء اﻟﺘﻌﺮف ﻋﻠﻰ ﻣﺴﺘﻮى اﻟﺠﻤﻠﺔ. اﻟﻨﻤﻮذج اﻟﻤﻘﺘﺮح ﺑﺄداء اﻷﻧﻈﻤﺔ اﻟﺤﺎﻟﯿﺔ اﻷﺧﺮى أن اﻟﻨﻤﻮذج اﻟﻤﻘﺘﺮح ﻗﺪ ﺗﻔﻮق ﻋﻠﻰ ھﺬه اﻷﻧﻈﻤﺔ. ﺛﺎﻧﯿﺎً، ﺗﻢ اﻗﺘﺮاح ﻧﻤﻮذج ﻣﺴﺘﻨﺪ إﻟﻰ اﻟﻤﺤﻮل .(Transformer) ﺗﻢ اﻻﺳﺘﻔﺎدة ﻣﻦ اﻻﻧﺘﺒﺎه اﻟﺬاﺗﻲ ﻣﻊ اﻻﺳﺘﻔﺎدة ﻣﻦ ﻣﺘﻌﺪد اﻟﺮؤوس اﻻﻧﺘﺒﺎه ﻟﺘﻄﻮﯾﺮ ﻧﻤﺎذج ﺷﺎﻣﻠﺔ ﻟﻠﺘﻌﺮف ﻋﻠﻰ اﻟﻜﺘﺎﺑﺔ اﻟﻌﺮﺑﯿﺔ. ﺗﺴﺘﺨﺪم اﻟﻤﺤﻮﻻت اّﻟﯿﺔ اﻻﻧﺘﺒﺎه اﻟﺬاﺗﻲ اﻟﺤﺴﺎﺑﯿﺔ ﺑﺸﻜﻞ ﻣﺘﻮازي واﻟﻘﺪرة ﻋﻠﻰ اﻟﺘﻘﺎط اﻟﺘﺒﻌﯿﺎت اﻟﻄﻮﯾﻠﺔ ﻓﻲ اﻟﺴﯿﺎﻗﺎت. اﻗﺘﺮﺣﻨﺎ ﻧﻤﻮذﺟﺎً ﻗﺎﺋﻤﺎً ﻋﻠﻰ اﻟﻤﺤﻮﻻت ﻷداء ﻣﮭﻤﺔ اﻟﺘﻌﺮف ﻋﻠﻰ ﺧﻂ اﻟﯿﺪ ﺷﺎﻣﻞ ﻟﻠﻜﺘﺎﺑﺔ اﻟﻌﺮﺑﯿﺔ ﺑﻄﺮﯾﻘﺔ ﺷﺎﻣﻠﺔ. ﻓﻲ ھﺬه اﻟﺪراﺳﺔ وﺿﺤﻨﺎ أن اﻟﻨﻤﻮذج اﻟﻤﻘﺘﺮح
There are no comments on this title.