Multiple-view D<sup>2</sup>NNs array: realizing robust 3D object recognition
Jiashuo Shi, Liang Zhou, Taige Liu, Chai Hu, Kewei Liu, Jun Luo, Haiwei Wang, Changsheng Xie, Xinyu Zhang
Abstract
As an optical-based classifier of the physical neural network, the independent diffractive deep neural network ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">D</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">N</mml:mi> </mml:mrow> </mml:math> ) can be utilized to learn the single-view spatial featured mapping between the input lightfields and the truth labels by preprocessing a large number of training samples. However, it is still not enough to approach or even reach a satisfactory classification accuracy on three-dimensional (3D) targets owing to already losing lots of effective lightfield information on other view fields. This Letter presents a multiple-view <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">D</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">s</mml:mi> </mml:mrow> </mml:math> array (MDA) scheme that provides a significant inference improvement compared with individual <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">D</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">N</mml:mi> </mml:mrow> </mml:math> or Res- <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">D</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">N</mml:mi> </mml:mrow> </mml:math> by constructing a different complementary mechanism and then merging all base learners of distinct views on an electronic computer. Furthermore, a robust multiple-view <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">D</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">N</mml:mi> <mml:mi mathvariant="normal">s</mml:mi> </mml:mrow> </mml:math> array (r-MDA) framework is demonstrated to resist the redundant spatial features of invalid lightfields due to severe optical disturbances.