eduzhai > Applied Sciences > Engineering >

Rotation Invariant Deep CBIR

  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: Introduction of Convolutional Neural Networks has improved results on almostevery image-based problem and Content-Based Image Retrieval is not anexception. But the CNN features, being rotation invariant, creates problems tobuild a rotation-invariant CBIR system. Though rotation-invariant features canbe hand-engineered, the retrieval accuracy is very low because by handengineering only low-level features can be created, unlike deep learning modelsthat create high-level features along with low-level features. This paper showsa novel method to build a rotational invariant CBIR system by introducing adeep learning orientation angle detection model along with the CBIR featureextraction model. This paper also highlights that this rotation invariant deepCBIR can retrieve images from a large dataset in real-time.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×