Object Detection using ORB Feature Descriptors and DBSCAN Clustering



Abstract:

This project aims to train a simple, traditional, non-cnn algorithm with ORB features of an object and DBSCAN clustering and test this algorithm on a set of test images. As opposed to above, here is a sample output of a CNN based approach:

CNN-based approach of object detection

Fig. 1: CNN-based approach of object detection




[TODO]