Bu kaydın yasal hükümlere uygun olmadığını düşünüyorsanız lütfen sayfa sonundaki Hata Bildir bağlantısını takip ederek bildirimde bulununuz. Kayıtlar ilgili üniversite yöneticileri tarafından eklenmektedir. Nadiren de olsa kayıtlarla ilgili hatalar oluşabilmektedir. MİTOS internet üzerindeki herhangi bir ödev sitesi değildir!

COUNTING NUMBER OF PEOPLE IN DIGITAL IMAGES USING FACE AND PEOPLE DETECTION ALGORITHMS

Oluşturulma Tarihi: 11-07-2016

Niteleme Bilgileri

Tür: Tez

Alt Tür: Yüksek Lisans Tezi

Yayınlanma Durumu: Yayınlanmamış

Dosya Biçimi: PDF

Dil: İngilizce

Konu(lar): TEKNOLOJİ, Teknik bilgiler iletişimi,

Yazar(lar): Husain, Samar Ittahir M.A (Yazar),

Emeği Geçen(ler): Koyuncu, Murat (Danışman),

Anahtar Kelimeler

 Counting People, Face Detection, People Detection, Skin Color, Viola Jones LBP, Viola Jones CART, Histogram of Oriented Gradients, Support Vector Machine.  


Özet

Counting the number of people in still images or video frames is an active research area that is a challenge in the computer vision field. It plays an important role in a variety of applications, such as security, management, education, and commerce. In this thesis, we work on counting the number of people in digital images. People can be seen differently in images, which requires the use of different techniques together. Therefore, we use two different techniques, which are Face Detection method and People Detection method, in order to estimate the number of people in an image. The proposed method combines the outputs of the Face Detection and People Detection methods in order to improve the performance of estimating the number of people in an input image with low cost and simple hardware. We test three face detection algorithms (Skin Color, Viola Jones LBP and Viola Jones CART) and a People Detection method (whole body) which is based on the HOG feature and SVM classifier to determine the best combination to estimate the number of people in input images. We use 240 test images including 1,202 people from two different datasets (Groups of Images of People and INRIA Person) to test the proposed system and determine the best combination. We have obtained best Recall of 91% and Precision of 93.97% by combining Viola Jones CART with People Detection method whereas by applying Face Detection and People Detection methods (whole body) separately, we got best Recall of 70.38% and Precision of 92.76% by Viola Jones CART method


İçindekiler



Açıklamalar



Haklar



Notlar



Kaynakça


Atıf Yapanlar

Gözat Sayfasına Dön

 

Sosyal Medya ve Araçlar

İstatistikler

  • Kayıt
    • Bu ay: 3
    • Toplam: 2417
  • Online
    • Ziyaretçi: 18
    • Üye: 0
    • Toplam: 18

Detaylı İstatistikler