Implementasi MPRL dan Hough Transform Untuk Segmentasi dan Ekstraksi Fitur Pada Citra Gaya Berjalan

Hustinawaty, Hustinawaty and Agushinta, Dewi and Kusuma, Tubagus Maulana and Lestari , Intan Nur and Damayanthi, Cicu Ratih (2011) Implementasi MPRL dan Hough Transform Untuk Segmentasi dan Ekstraksi Fitur Pada Citra Gaya Berjalan. In: Seminar Nasional dan Expo Teknik Elektro 2011.

Hustinawaty Seminar Nasional dan ExpoTeknik Elektro 2011.pdf

Download (605Kb) | Preview


Human gait is unique. It has different characteristics and have a high ability to distinguish a person. "Gait” analysis, mapping human body points is todays advanced technology. Stages of "gait" processing start from segmentation, extraction until recognition. This paper discusses the segmentation and extraction of gait features. Thinning methods with Most Prominent Ridge Line (MPRL) algorithms to form the skeleton bones of human beings is used. The resulting skeleton image is used for features gait extraction through annular analysis. Method for feature extraction used is Hough transformation. The resulting features are grouped according to distance and angle. This grouping determines the location of features on an image skeleton. The next process is to calculate the distance between feature and angle formed. There are ten features obtained. Four features of a group spacing distance of the front foot until the heel of the hind legs, hands up to the front axle center of the chest, arms rear axle until the middle of the chest and arms front to the back of the hand. Six other features are the result of measurement angle of the front knee, ankle front, rear ankle, knee hind legs, forearms and elbows behind the elbow sleeves.

Item Type: Conference or Workshop Item (Paper)
Subjects: T Technology > T Technology (General)
Divisions: Fakultas Ilmu Komputer dan Teknologi Informasi > Program Studi Sistem Informasi
Depositing User: Mr Reza Chandra
Date Deposited: 25 May 2016 10:05
Last Modified: 25 May 2016 10:05

Actions (login required)

View Item View Item