Object Recognition for a Blind Relief System
Object recognition refers to identification of a specific object in an image or video. It can have immense applications like surveillance, navigation, etc. Another application could be an aid for the blind. Object recognition can help a blind person in his daily routine to identify any product or object that comes in front of him, maybe when he is in a supermarket. This work proposes a method of object recognition in real-time and give a speech output of the object or product present in the field of view of a camera by comparing it with images in the dataset. In the proposed system, a query image is categorized using the k-nearest neighbor classifier that makes use of Scale Invariant Feature Transform (SIFT). SIFT features are scale, orientation and illumination invariant. These features of the images are extracted to form a similarity matrix by matching these SIFT features. We then implement the k-nearest neighbor classifier of this similarity matrix.
Keywords - Scale Invariant Feature Transform (SIFT), classifier, k-NN Classifier