Al-Smadi M.Abdulrahim K.Salama R.A.2024-05-282024-05-282019227738782-s2.0-85067007474https://www.scopus.com/inward/record.uri?eid=2-s2.0-85067007474&partnerID=40&md5=d47e75e4c40e5c0235f48740421e5e69https://www.ijrte.org/wp-content/uploads/papers/v7i6s2/F10490476S219.pdfhttps://oarep.usim.edu.my/handle/123456789/8902Accurate and precise vehicle recognition and classification play a major role in analyzing and understanding traffic surveillance systems. This paper proposes a dynamic feature descriptor to recognize and classify road users based on graph representation. Local gradient patterns are computed based on the grayscale difference on the four directions across the center pixels. Dynamic gradients are determined according to the effective gradient computed as the mean value of all gradients. Hierarchal Graph using angular rotation pattern are applied to extract Dynamic Gradient Patterns (DGP). The central pixel is represented by Hierarchal Graph of Dynamic Gradient Patterns (HG-DGP).The proposed method learns dynamic representation adaptively to achieve efficient recognition with higher accuracy and lower pre-processing. The experimental results show that the proposed technique combined with support vector machine is efficient and discriminative for road user recognition and classification. � BEIESP.en-USArticle41441876