Biologically Inspired Tensor Features

According to the research results reported in the past decades, it is well acknowledged that face recognition is not a trivial task. With the development of electronic devices, we are gradually revealing the secret of object recognition in the primate's visual cortex. Therefore, it is time to reconsider face recognition by using biologically inspired features. In this paper, we represent face images by utilizing the C1 units, which correspond to complex cells in the visual cortex, and pool over S1 units by using a maximum operation to reserve only the maximum response of each local area of S1 units. The new representation is termed C1 Face. Because C1 Face is naturally a third-order tensor (or a three dimensional array), we propose three-way discriminative locality alignment (TWDLA), an extension of the discriminative locality alignment, which is a top-level discriminate manifold learning-based subspace learning algorithm. TWDLA has the following advantages: (1) it takes third-order tensors as input directly so the structure information can be well preserved; (2) it models the local geometry over every modality of the input tensors so the spatial relations of input tensors within a class can be preserved; (3) it maximizes the margin between a tensor and tensors from other classes over each modality so it performs well for recognition tasks and (4) it has no under sampling problem. Extensive experiments on YALE and FERET datasets show (1) the proposed C1Face representation can better represent face images than raw pixels and (2) TWDLA can duly preserve both the local geometry and the discriminative information over every modality for recognition.

A voltage-controlled tunable two-color infrared photodetector using GaAs/AlAs/GaAlAs and GaAs/GaAlAs stacked multiquantum wells

A voltage-controlled tunable two-color infrared detector with photovoltaic (PV) and photoconductive (PC) dual-mode operation at 3-5 mu m and 8-14 mu m using GaAs/AlAs/AlGaAs double barrier quantum wells (DBQWs) and bound-to-continuum GaAs/AlGaAs quantum wells is demonstrated. The photoresponse peak of the photovoltaic GaAs/AlAs/GaAlAs DBQWs is at 5.3 mu m, and that of the photoconductive GaAs/GaAlAs quantum wells is at 9.0 mu m. When the two-color detector is under a zero bias, the spectral response at 5.3 mu m is close to saturate and the peak detectivity at 80 K can reach 1.0X10(11) cmHz(1/2)/W, while the spectral photoresponsivity at 9.0 mu m is absolutely zero completely. When the external voltage of the two-color detector is changed to 2.0 V, the spectral photoresponsivity at 5.3 mu m becomes zero while the spectral photoresponsivity at 9.0 mu m increases comparable to that at 5.3 mu m under zero bias, and the peak detectivity (9.0 mu m) at 80 K can reach 1.5X10(10) cmHz(1/2)/W. Strictly speaking, this is a real bias-controlled tunable two-color infrared photodetector. We have proposed a model based on the PV and PC dual-mode operation of stacked two-color QWIPs and the effects of tunneling resonance with narrow energy width of photoexcited electrons in DBQWs, which can explain qualitatively the voltage-controlled tunable behavior of the photoresponse of the two-color infrared photodetector. (C) 1996 American Institute of Physics.