Title :
An efficient Naive Bayes approach to category-level object detection
Author :
Terzic, Kasim ; du Buf, J.M.H.
Author_Institution :
Vision Lab. (LARSyS), Univ. of the Algarve, Faro, Portugal
Abstract :
We present a fast Bayesian algorithm for category-level object detection in natural images. We modify the popular Naive Bayes Nearest Neighbour classification algorithm to make it suitable for evaluating multiple sub-regions in an image, and offer a fast, filtering-based alternative to the multi-scale sliding window approach. Our algorithm is example-based and requires no learning. Tests on standard datasets and robotic scenarios show competitive detection rates and real-time performance of our algorithm.
Keywords :
Bayes methods; filtering theory; image classification; object detection; category-level object detection; competitive detection rates; fast filtering-based approach; multiple sub-region evaluation; multiscale sliding window approach; naive Bayes nearest neighbour classification algorithm; natural images; robotic scenarios; Complexity theory; Kernel; Object detection; Real-time systems; Robots; Testing; Training; Computer vision; Nearest neighbour; Object detection; Real time systems; Robot vision;
Conference_Titel :
Image Processing (ICIP), 2014 IEEE International Conference on
Conference_Location :
Paris
DOI :
10.1109/ICIP.2014.7025332