<?xml version="1.0" encoding="UTF-8"?>
<collection xmlns="http://www.loc.gov/MARC21/slim">
 <record>
  <leader>02854nam a2200217Ia 4500</leader>
  <controlfield tag="001">CTU_233380</controlfield>
  <controlfield tag="008">210402s9999    xx            000 0 und d</controlfield>
  <datafield tag="082" ind1=" " ind2=" ">
   <subfield code="a">006.425</subfield>
  </datafield>
  <datafield tag="082" ind1=" " ind2=" ">
   <subfield code="b">Ch312</subfield>
  </datafield>
  <datafield tag="088" ind1=" " ind2=" ">
   <subfield code="a">8480104</subfield>
  </datafield>
  <datafield tag="100" ind1=" " ind2=" ">
   <subfield code="a">Trần, Văn Út Chính</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2="0">
   <subfield code="a">Sử dụng mạng Neural tích chập để nhận dạng Ký tự viết tay chữ Hoa tiếng Việt :</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2="0">
   <subfield code="b">Luận văn thạc sĩ ngành. Công nghệ Thông tin</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2="0">
   <subfield code="c">Trần Văn Út Chính ; Trần Cao Đệ (Hướng dẫn khoa học)</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
   <subfield code="a">Cần Thơ</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
   <subfield code="b">Trường Đại học Cần Thơ</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
   <subfield code="c">2019</subfield>
  </datafield>
  <datafield tag="520" ind1=" " ind2=" ">
   <subfield code="a">Nhận dạng chữ viết tay là một lĩnh vực nghiên cứu tích cực về trí tuệ nhân tạo, nhận dạng mẫu và thị giác máy tính. Lĩnh vực nhận dạng văn bản đã đạt được thành công lớn trong các ứng dụng thực tiễn, đặc biệt là trong hệ thống chính phủ điện tử, ứng dụng bảo mật và các lĩnh vực khác. Các hệ thống nhận dạng ký tự viết tay tiếng Việt phải đối mặt với một số thách thức do sự thay đổi phong cách trong chữ viết tay của con người và chưa có bộ cơ sở dữ liệu chuẩn. Trong luận văn này, mô hình hóa một kiến trúc học tập sâu được áp dụng hiệu quả để nhận ra các ký tự viết tay chữ hoa tiếng Việt. Mạng nơ ron tích chập (CNN) là một loại đa lớp chuyển tiếp đặc biệt được huấn luyện ở chế độ giám sát. Mô hình CNN đã huấn luyện và thử nghiệm cơ sở dữ liệu của chúng tôi có chứa 17.800 ký tự viết tay chữ hoa tiếng Việt. Trong luận văn này, các phương pháp tối ưu hóa được triển khai để tăng hiệu suất của CNN. Các phương pháp máy học thông thường thường áp dụng kết hợp trình trích xuất tính năng và trình phân loại có thể huấn luyện. Việc sử dụng CNN đưa đến kết quả nhận dạng được cải tiến đáng kể so với một số giải pháp dùng thuật toán phân loại máy học khác. Mô hình CNN được đề xuất với kết quả 97% trên dữ liệu kiểm tra.</subfield>
  </datafield>
  <datafield tag="650" ind1=" " ind2=" ">
   <subfield code="a">Pattern recognition systems,Hệ thống nhận dạng mẫu</subfield>
  </datafield>
  <datafield tag="910" ind1=" " ind2=" ">
   <subfield code="c">tvtrong</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
   <subfield code="a">Trung tâm Học liệu Trường Đại học Cần Thơ</subfield>
  </datafield>
 </record>
</collection>
