Object recognition for blind people using portable camera

For blind persons it is very essential to recognize a product of their daily use so we implied a method to identify product in their everyday routine by use of camera. To separate a object from un-necessary background, movement based technique is used to spot object of concern from the video by instructing person to shake the object that gives moving object area by K-means clustering background separation technique. After detection of moving area feature extraction SIFT technique is used to take out key points of that product and that key-point is matched with the data base images if it matches efficiently then it identify a product and produce name of that product to blind persons in speech. Instead of OCR we used SIFT because in OCR orientation of camera must be towards label of product and for blind person it is very tough to locate label efficiently whereas in SIFT product orientation is not a big deal as key-points can be matched from any position.