Embedding Rotate-and-Scale Net for Learning Invariant Features of Simple Images

2019 ◽  
pp. 131-142
Author(s):  
Zihang He ◽  
Xiang Ye ◽  
Zuguo He ◽  
Yong Li
Keyword(s):  
Invariant Features

Image quality assessment based on local invariant features

2013 ◽  
Vol 32 (12) ◽  
pp. 3369-3372 ◽  
Author(s):  
Ya-zhou YANG ◽  
Xiao-qing YING ◽  
Guang-quan CHENG ◽  
Dan TU
Keyword(s):  
Image Quality ◽  
Quality Assessment ◽  
Image Quality Assessment ◽  
Local Invariant ◽  
Invariant Features ◽  
Local Invariant Features

Secure Watermarking using Diophantine Equations for Authentication and Recovery

2015 ◽  
Vol 3 (2) ◽  
Author(s):  
Jayashree Nair ◽  
T. Padma
Keyword(s):  
Diophantine Equation ◽  
Diophantine Equations ◽  
Jpeg Compression ◽  
Authentication Scheme ◽  
Original Image ◽  
Invariant Features ◽  
Jpeg2000 Compression ◽  
Frequency Properties ◽  
Visual Authentication

This paper describes an authentication scheme that uses Diophantine equations based generation of the secret locations to embed the authentication and recovery watermark in the DWT sub-bands. The security lies in the difficulty of finding a solution to the Diophantine equation. The scheme uses the content invariant features of the image as a self-authenticating watermark and a quantized down sampled approximation of the original image as a recovery watermark for visual authentication, both embedded securely using secret locations generated from solution of the Diophantine equations formed from the PQ sequences. The scheme is mildly robust to Jpeg compression and highly robust to Jpeg2000 compression. The scheme also ensures highly imperceptible watermarked images as the spatio –frequency properties of DWT are utilized to embed the dual watermarks.

scholarly journals Learning Invariant Features Using Subspace Restricted Boltzmann Machine

2016 ◽  
Vol 45 (1) ◽  
pp. 173-182 ◽  
Author(s):  
Jakub M. Tomczak ◽  
Adam Gonczarek
Keyword(s):  
Restricted Boltzmann Machine ◽  
Boltzmann Machine ◽  
Invariant Features

Fuzzy ARTMAP classification of invariant features derived using angle of rotation from a neural network

2000 ◽  
Vol 130 (1-4) ◽  
pp. 67-84 ◽  
Author(s):  
P. Raveendran ◽  
R. Palaniappan ◽  
Sigeru Omatu
Keyword(s):  
Neural Network ◽  
Fuzzy Artmap ◽  
Invariant Features ◽  
Angle Of Rotation

scholarly journals Topological Sensitivity in the Recognition of Disoriented Figures

i-Perception ◽  
2018 ◽  
Vol 9 (6) ◽  
pp. 204166951880971 ◽  
Author(s):  
Fumio Kanbe
Keyword(s):  
Structural Properties ◽  
Topological Structure ◽  
Topological Property ◽  
Topological Sensitivity ◽  
Topological Structures ◽  
Invariant Features

A previous study by the author found that discrimination latencies for figure pairs with the same topological structure (isomorphic pairs) were longer than for pairs with different topological structures (nonisomorphic pairs). These results suggest that topological sensitivity occurs during figure recognition. However, sameness was judged in terms of both shape and orientation. Using this criterion, faster discrimination of nonisomorphic pairs may have arisen from the detection of differences in the corresponding locations of the paired figures, which is not a topological property. The current study examined whether topological sensitivity occurs even when identity judgments are based on the sameness of shapes, irrespective of their orientation, where the sameness of location is not ensured. The current results suggested the involvement of topological sensitivity, indicating that processing of structural properties (invariant features) of a figure may be prioritized over processing of superficial features, such as location, length, and angles, in figure recognition.

Sign in / Sign up

Export Citation Format

Share Document