Detection and Recognition of Outdoor Objects based on SIFT Features

Abstract: Object detection and recognition is very essential for visually impaired to survive at indoor and outdoor environments. The objective of this paper is to detect and recognize the outdoor objects. In an effective method is proposed and implemented for the extraction of the objects from the outdoor environment automatically. The outdoor objects are detected from the real-time outdoor images. The outdoor objects are detected using Haar-like features and AdaBoost classifier. SIFT features are extracted from the detected objects and classified using Support Vector Machine. The experimental results reveals that the detection and recognition rate for real-time outdoor object using SIFT with SVM is 91.10%.

Keywords: object detection, AdaBoost Classifier, Haar-like feature, Support Vector Machine (SVM) classifier, Scale Invariant Feature Transform (SIFT).