arXiv subjects

Xin Lei

Publications and source records attributed to Xin Lei.

23 records · Page 2Linked to original sources

arXiv2019

Direct Object Recognition Without Line-of-Sight Using Optical Coherence

Visual object recognition under situations in which the direct line-of-sight is blocked, such as when it is occluded around the corner, is of practical importance in a wide range of applications. With coherent illumination, the light scattered from diffusive walls forms speckle patterns that contain information of the hidden object. It is possible to realize non-line-of-sight (NLOS) recognition with these speckle patterns. We introduce a novel approach based on speckle pattern recognition with deep neural network, which is simpler and more robust than other NLOS recognition methods. Simulations and experiments are performed to verify the feasibility and performance of this approach.

cs.CV↗

arXiv2019

End-To-End Speech Recognition Using A High Rank LSTM-CTC Based Model

Long Short Term Memory Connectionist Temporal Classification (LSTM-CTC) based end-to-end models are widely used in speech recognition due to its simplicity in training and efficiency in decoding. In conventional LSTM-CTC based models, a bottleneck projection matrix maps the hidden feature vectors obtained from LSTM to softmax output layer. In this paper, we propose to use a high rank projection layer to replace the projection matrix. The output from the high rank projection layer is a weighted combination of vectors that are projected from the hidden feature vectors via different projection matrices and non-linear activation function. The high rank projection layer is able to improve the expressiveness of LSTM-CTC models. The experimental results show that on Wall Street Journal (WSJ) corpus and LibriSpeech data set, the proposed method achieves 4%-6% relative word error rate (WER) reduction over the baseline CTC system. They outperform other published CTC based end-to-end (E2E) models under the condition that no external data or data augmentation is applied. Code has been made available at https://github.com/mobvoi/lstm_ctc.

cs.CL↗

arXiv2019

Imaging-free object recognition enabled by optical coherence

Visual object recognition is one of the most important perception functions for a wide range of intelligent machines. A conventional recognition process begins with forming a clear optical image of the object, followed by its computer analysis. In contrast, it is possible to carry out recognition without imaging by using coherent illumination and directly analyzing the optical interference pattern of the scattered light as captured by an image sensor. Here we show that such direct visual recognition can overcome traditional limitations of imaging optics to realize excellent recognition without focusing, beyond diffraction limit, or in the absence of direct line-of-sight.

eess.IV↗

arXiv2017

A non-oscillatory energy-splitting method for the computation of compressible multi-fluid flows

This paper proposes a new non-oscillatory {\em energy-splitting} conservative algorithm for computing multi-fluid flows in the Eulerian framework. In comparison with existing multi-fluid algorithms in literatures, it is shown that the mass fraction model with isobaric hypothesis is a plausible choice for designing numerical methods for multi-fluid flows. Then we construct a conservative Godunov-based scheme with the high order accurate extension by using the generalized Riemann problem (GRP) solver, through the detailed analysis of kinetic energy exchange when fluids are mixed under the hypothesis of isobaric equilibrium. Numerical experiments are carried out for the shock-interface interaction and shock-bubble interaction problems, which display the excellent performance of this type of schemes and demonstrate that nonphysical oscillations are suppressed around material interfaces substantially.

physics.flu-dyn↗

arXiv2017

Explicit three dimensional topology optimization via Moving Morphable Void (MMV) approach

Three dimensional (3D) topology optimization problems always involve huge numbers of Degrees of Freedom (DOFs) in finite element analysis (FEA) and design variables in numerical optimization, respectively. This will inevitably lead to large computational efforts in the solution process. In the present paper, an efficient and explicit topology optimization approach which can reduce not only the number of design variables but also the number of degrees of freedom in FEA is proposed based on the Moving Morphable Voids (MMVs) solution framework. This is achieved by introducing a set of geometry parameters (e.g., control points of B-spline surfaces) to describe the boundary of a structure explicitly and removing the unnecessary DOFs from the FE model at every step of numerical optimization. Numerical examples demonstrate that the proposed approach does can overcome the bottleneck problems associated with a 3D topology optimization problem in a straightforward way and enhance the solution efficiency significantly.

cond-mat.other↗