A study on video based childand adult classification with biometry
Başlık çevirisi mevcut değil.
- Tez No: 707405
- Danışmanlar: DR. JONGKWAN SONG, DR. JANG-SİK PARK
- Tez Türü: Yüksek Lisans
- Konular: Biyomühendislik, Biyoteknoloji, Elektrik ve Elektronik Mühendisliği, Bioengineering, Biotechnology, Electrical and Electronics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2014
- Dil: İngilizce
- Üniversite: Kyungsung University
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 69
Özet
Özet yok.
Özet (Çeviri)
In this day and time, number of social crimes is getting bigger. However one of them may threaten the future of the world more. It is child abduction. If children cannot live good childhood with good conditions, it would be pointless to expect children to be role model. For this reason, something has to be done. We need to decrease the number of crimes, and make a contribution to social security. Since image processing can be useful and help to provide social security, it can be used for the benefits of society. Pedestrian detection is popular topic currently and there are various studies on image and video based pedestrian detection. However, there is not many study on detection and classification according to people's ages. By unifying pedestrian detection idea with age classification, so many applications can be made for the benefit of people. In this study, a new algorithm for child and adult classification is proposed. This algorithm is based on to calculate biometry by using cascade classifiers which are so popular on image and video processing currently. In this study, I used the full body and head and shoulder classifiers to detect each of them respectively. Since I deal with classifying people according to their ages, biometry is so deterministic. Making relative measurement is the main point of this study. The reason is that head size of people gets bigger before they become teenager, therefore children look so adorable. The first step of the algorithm is detecting pedestrian's full body as a rectangle in the video, and then, next step is finding head area inside the full body rectangle. In the last step, the only thing we need to do is proportioning length of head rectangle over length of full body rectangle. In this way, we can see what our ratio is, and next, we compare that ratio to threshold value. If ratio is smaller than the threshold value, that means, the person we detected is adult. Otherwise it's a child. Experimental results showed that the proposed algorithm can classify people properly according to their ages. And this results give a big impression for improving this work up to higher percentages.
Benzer Tezler
- Zihin yetersizliği olan çocuğun oyun becerilerinin geliştirilmesinde Doğal Öğretim Temelli Aile Eğitim Programının (DOTAP) incelenmesi: Vaka çalışması
Investigation of Naturalistic Teaching-Based Family Education Program (DOTAP) in the development of play skills of children with intellectual disabilities: A case study
AYŞE KAYA
Yüksek Lisans
Türkçe
2023
Eğitim ve ÖğretimEskişehir Osmangazi ÜniversitesiÖzel Eğitim Ana Bilim Dalı
DR. ÖĞR. ÜYESİ MERAL MELEKOĞLU
- Pediatrik kronik böbrek hastalarında sanal gerçeklik tabanlı ve video destekli egzersiz uygulamalarının etkinliğinin karşılaştırılması
Comparison of the efficiency of virtual reality-based and video-supported exercises in pediatric chronic kidney disease
IRMAK ÇAVUŞOĞLU
Doktora
Türkçe
2022
Fizyoterapi ve Rehabilitasyonİstanbul Medipol ÜniversitesiFizyoterapi ve Rehabilitasyon Ana Bilim Dalı
DOÇ. DR. DEVRİM TARAKCI
- Okul öncesi çağda işitme engelli çocuğu olan normal işiten bir annenin grup oyunu esnasında kullandığı stratejilerin incelenmesi
Başlık çevirisi yok
NURHAYAT COŞKUN
- Video based automatic affect analysis of a child and a therapist during play therapy
Bir çocuk ve bir terapistin oyun terapisi sırasındaki duygu durumunun videodan otomatik analizi
BATIKAN TÜRKMEN
Yüksek Lisans
İngilizce
2019
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolBoğaziçi ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
PROF. DR. LALE AKARUN ERSOY
DOÇ. DR. ALBERT ALİ SALAH