Recognizing Objects Using Screening Descriptors

Question

I am new to this area. My problem is to determine if the object is present in the object used in the training images (images of similar objects) on the test image or not. I want to use SIFT descriptors for recognition. Is the word bag suitable by clustering SIFT descriptors, which is used to classify objects into different classes, suitable for it, or if there is a simpler approach using sift descriptors for it. thanks in advance

Answer 1

The Bag of Visual Words (BoW) is truly a classic approach, originally proposed by Sivich and Sisserman in 2003 [ Paper ]. He was one of the first to deviate from previous methods that preferred global descriptors over local functions such as SIFT and SURF. I recommend continuing with the implementation of this classic pipeline if you are just starting to learn about object detection and recognition.