The ability to form images through hair-thin optical fibres promises to open up new applications from biomedical imaging to industrial inspection. Unfortunately, their deployment has been limited because small changes in mechanical deformation (e.g. bending) and temperature can completely scramble optical information, which distorts the resulting images. Since such changes are dynamic, correcting them requires measurement of the fibre transmission matrix in situ immediately before imaging. Transmission matrix calibration typically requires access to both the proximal and distal facets of the fibre simultaneously, which is not feasible during most realistic usage scenarios without compromising the thin form factor with bulky distal optics. Here, we introduce a new approach to determine the transmission matrix of multi-mode or multi-core optical fibre in a reflection-mode configuration without requiring access to the distal facet. A thin stack of structured metasurface reflectors is used at the distal facet of the fibre to introduce wavelength-dependent, spatially heterogeneous reflectance profiles. We derive a first-order fibre model that compensates these wavelength-dependent changes in the fibre transmission matrix and show that, consequently, the reflected data at 3 wavelengths can be used to unambiguously reconstruct the full transmission matrix by an iterative optimisation algorithm. We then present a method for sample illumination and imaging following reconstruction of the transmission matrix. Unlike previous approaches, our method does not require the fibre matrix to be unitary making it applicable to physically realistic fibre systems that have non-negligible power loss. We demonstrate the transmission matrix reconstruction and imaging method first using simulated non-unitary fibres and noisy reflection matrices, then using much larger experimentally-measured transmission matrices of a densely-packed multicore fibre. Finally, we demonstrate the method on an experimentally-measured multi-wavelength set of transmission matrices recorded from a step-index multimode fibre. Our findings pave the way for online transmission matrix calibration in situ in hair-thin optical fibres.
LPCNet is an efficient vocoder that combines linear prediction and deep neural network modules to keep the computational complexity low. In this work, we present two techniques to further reduce it's complexity, aiming for a low-cost LPC-Net vocoder-based neural Text-to-Speech (TTS) System. These techniques are: 1) Sample-bunching, which allows LPCNet to generate more than one audio sample per inference; and 2) Bitbunching, which reduces the computations in the final layer of LPCNet. With the proposed bunching techniques, LPCNet, in conjunction with a Deep Convolutional TTS (DCTTS) acoustic model, shows a 2.19x improvement over the baseline run-time when running on a mobile device, with a less than 0.1 decrease in TTS mean opinion score (MOS).
We address the numerical challenge of solving regular Sturm-Liouville problems in Liouville's normal form, with a continuous and piecewise analytic potential and self-adjoint separated boundary conditions. The novelty of our approach, which is based on a non-standard truncation of Fer expansions, which we call 'Fer streamers', lies in the construction of a new numerical method, which, i) does not impose any restriction on the step size for eigenvalues which are greater than or equal to the minimum of the potential, ii) requires only a mild restriction on the step size for the remaining finite number of eigenvalues, iii) can attain any convergence rate, which grows exponentially with the number of terms, and is uniform for every eigenvalue, and, iv) lends itself to a clear understanding of the manner in which the potential affects the local and global errors. We provide our numerical method with its analytical underpinning, but emphasize that it is at an early stage of development and that much remains to be done. In particular, we comment on our investigation of efficient discretization schemes for the integrals which arise in Fer streamers.
We introduce a framework for the reconstruction of the amplitude, phase, and polarization of an optical vector-field using measurements acquired by an imaging device characterized by an integral transform with an unknown spatially variant kernel. By incorporating effective regularization terms, this new approach is able to recover an optical vector-field with respect to an arbitrary representation system, which may be different from the one used for device calibration. In particular, it enables the recovery of an optical vector-field with respect to a Fourier basis, which is shown to yield indicative features of increased scattering associated with tissue abnormalities. We demonstrate the effectiveness of our approach using synthetic holographic images and biological tissue samples in an experimental setting, where the measurements of an optical vector-field are acquired by a multicore fiber endoscope, and observe that indeed the recovered Fourier coefficients are useful in distinguishing healthy tissues from tumors in early stages of oesophageal cancer.
Increasing demand for on-device Automatic Speech Recognition (ASR) systems has resulted in renewed interests in developing automatic model compression techniques. Past research have shown that AutoML-based Low Rank Factorization (LRF) technique, when applied to an end-to-end Encoder-Attention-Decoder style ASR model, can achieve a speedup of up to 3.7×, outperforming laborious manual rank-selection approaches. However, we show that current AutoML-based search techniques only work up to a certain compression level, beyond which they fail to produce compressed models with acceptable word error rates (WER). In this work, we propose an iterative AutoML-based LRF approach that achieves over 5× compression without degrading the WER, thereby advancing the state-of-the-art in ASR compression.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.