Designing Deep Neural Networks (DNNs) running on edge hardware remains a challenge. Standard desi... more Designing Deep Neural Networks (DNNs) running on edge hardware remains a challenge. Standard designs have been adopted by the community to facilitate the deployment of Neural Network models. However, not much emphasis is put on adapting the network topology to fit hardware constraints. In this paper, we adapt one of the most widely used architectures for mobile hardware platforms, MobileNetV2, and study the impact of changing its topology and applying post-training quantization. We discuss the impact of the adaptations and the deployment of the model on an embedded hardware platform for face detection.
The Aeronautics industry has pioneered safety from digital checklists to moving maps that improve... more The Aeronautics industry has pioneered safety from digital checklists to moving maps that improve pilot situational awareness and support safe ground movements. Today, pilots deal with increasingly complex cockpit environments and air traffic densification. Here we present an intelligent vision system, which allows real-time human-machine interaction in the cockpits to reduce pilot’s workload. The challenges for such a vision system include extreme change in background light intensity, large field-of-view and variable working distances. Adapted hardware, use of state-of-the-art computer vision techniques and machine learning algorithms in eye gaze detection allow a smooth, and accurate real-time feedback system. The current system has been over-specified to explore the optimized solutions for different use-cases. The algorithmic pipeline for eye gaze tracking was developed and iteratively optimized to obtain the speed and accuracy required for the aviation use cases. The pipeline, which is a combination of data-driven and analytics approaches, runs in real time at 60 fps with a latency of about 32ms. The eye gaze estimation error was evaluated in terms of the point of regard distance error with respect to the 3D point location. An average error of less than 1.1cm was achieved over 28 gaze points representing the cockpit instruments placed at about 80-110cm from the participants’ eyes. The angular gaze deviation goes down to less than 1° for the panels towards which an accurate eye gaze was required according to the use cases.
We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in... more We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in thin gold films. The hole arrays are fabricated by a low-cost technique using convectively self-assembled polystyrene beads as an initial pattern template. Our results suggest that corrugated, short-range ordered hole arrays allow the excitation of resonant surface plasmons. This opens the door to inexpensive, largescale and high throughput fabrication of nanophotonic devices.
A photonic crystal structure for condensing electromagnetic radiation comprising a modulated squa... more A photonic crystal structure for condensing electromagnetic radiation comprising a modulated square-shaped lattice and a corresponding condensing device are disclosed
Multispectral imagers collect a hypercube of data, where the spatial image is along two-dimension... more Multispectral imagers collect a hypercube of data, where the spatial image is along two-dimensions and the spectral information is in the third. Two main technologies are used for multispectral imaging: sweeping, where the hypercube is built by scanning through different wavelengths or spatial positions and snapshot multispectral spectral imaging, where the 3D cube of images is taken in one shot. Sweeping imaging systems tend to have more lines and better spectral resolutions whilst snapshot cameras are often used for dynamic analysis of scenes. A common method to obtain the hypercube in snapshot imagers is by pixel level filtering on the sensor chip. Pixel level filtering, where the filter is placed directly on the pixels are intergrated into the wafer-level making processing making them difficult to customize. Therefore, these sensors tend to aim for equally spaced spectral lines in order to cover many applications. This results in an often in an unnecessarily large data cube when...
We have designed, fabricated and tested narrow-band Fabry-Perot filters in the infrared using gol... more We have designed, fabricated and tested narrow-band Fabry-Perot filters in the infrared using gold porous mirrors and a silicon spacer layer. The filter peaks at 10 μm and 15 μm have approximately 10% transmission and a 1.5% linewidth. A Fabry-Perot structure with plane metal layers having a similar linewidth would have a transmission of only 0.2%. Thus, for the same linewidth we have improved the transmission by a factor of 50. Apart from the optical enhancements, these filters also have the advantage that they can be made inexpensively in a standard silicon MEMS technology and that their resonances can be finely tuned through post processing.
In this paper, we experimentally studied both the bright-field and fluorescence images of microsp... more In this paper, we experimentally studied both the bright-field and fluorescence images of microspheres by conventional and confocal scanning polarization microscopes. A qualitative analysis have been given to show a physical picture on the imaging of the microspheres. Emission spectra from melamine formaldehyde microspheres stained with Ethidium Bromide or covered by thin shell of CdTe nanocrystals have been experimentally studied. We adopted analytical expressions describing the resonance spacing in order to determine the size of the microspheres.
Confocal, Multiphoton, and Nonlinear Microscopic Imaging II, 2005
Confocal microscopy has a unique optical sectioning property which allows three-dimensional image... more Confocal microscopy has a unique optical sectioning property which allows three-dimensional images at different depths. Use of a microlens array is a potential alternative to the Nipkow disk for parallel imaging with high throughput in real-time confocal microscopy. The use of variable-focal-length microlenses can provide a way to axially scan the foci electronically avoiding the inflexible mechanical movement of the
The use of variable-focal-length (VFL) microlenses can provide a way to axially scan the foci acr... more The use of variable-focal-length (VFL) microlenses can provide a way to axially scan the foci across a sample by electronic control. We demonstrate an approach to coupling VFL microlenses individually to a fiber bundle as a way to create a high-throughput aperture array with a controllable aperture pattern. It would potentially be applied in real-time confocal imaging in vivo for biological specimens. The VFL microlenses that we used consist of a liquid-crystal film sandwiched between a pair of conductive substrates for which one has a hole-patterned electrode. One obtains the variation of the focal length by changing the applied voltage. The fiber bundle has been characterized by coupling with both coherent and incoherent light sources. We further demonstrate the use of a VFL microlens array in combination with the fiber bundle to build up a confocal system. The axial response of the confocal system has been measured without mechanical movement of the sample or the objective, and the FWHM is estimated to be approximately 16 m, with asymmetric sidelobes.
2020 14th International Symposium on Medical Information Communication Technology (ISMICT), 2020
Monitoring of breathing rate considerably improves healthcare quality by providing information ab... more Monitoring of breathing rate considerably improves healthcare quality by providing information about the patient’s state. Yet, the current methods used to assess the respiratory rate have some drawbacks that could be improved. Recent developments of the non-contact based measurement of the respiratory rate brought the remote Photoplethysmography (rPPG) technique. This technique allows physiological parameters monitoring with a camera and ambient light. It measures the light absorption variation by the blood to extract several physiological parameters. The main issue with the rPPG method is that its signal quality is lower compared with the contact-based methods. Several ways to improve this quality may be explored to obtain a robust rPPG pipeline. One of these is the combination of the multidimensional temporal traces generated by a camera into a single temporal trace. The literature methods aimed to extract a signal with the maximal information using smart estimation of the combination coefficients. These coefficients are computed using statistical or physiological/optical properties of the traces. In this paper, we developed an algorithm that estimates a better respiratory signal than with regular methods. The algorithm named Energy Variance Maximization (EVM) estimates the linear combination that maximizes the Signal Noise Ratio (SNR) of the output trace. We compare our contribution with three state-of-the-art method: CHROM, PBV and PVM. The results obtained with this method are better than the state of the art methods, with the Mean Absolute Error (MAE) being 0.87 rpm better than CHROM, the best of compared literature methods.
The problem of estimating a surface shape from its observed reflectance properties still remains ... more The problem of estimating a surface shape from its observed reflectance properties still remains a challenging task in computer vision. The presence of global illumination effects such as inter-reflections or cast shadows makes the task particularly difficult for non-convex real-world surfaces. State-of-the-art methods for calibrated photometric stereo address these issues using convolutional neural networks (CNNs) that primarily aim to capture either the spatial context among adjacent pixels or the photometric one formed by illuminating a sample from adjacent directions. In this paper, we bridge these two objectives and introduce an efficient fully-convolutional architecture that can leverage both spatial and photometric context simultaneously. In contrast to existing approaches that rely on standard 2D CNNs and regress directly to surface normals, we argue that using separable 4D convolutions and regressing to 2D Gaussian heat-maps severely reduces the size of the network and make...
2020 7th Swiss Conference on Data Science (SDS), 2020
Image denoising is an essential tool in computational photography. Standard denoising techniques,... more Image denoising is an essential tool in computational photography. Standard denoising techniques, which use deep neural networks at their core, require pairs of clean and noisy images for its training. If we do not possess the clean samples, we can use blind-spot neural network architectures, which estimate the pixel value based on the neighbouring pixels only. These networks thus allow training on noisy images directly, as they by-design avoid trivial solutions. Nowadays, the blind-spot is mostly achieved using shifted convolutions or serialization. We propose a novel fully convolutional network architecture that uses dilations to achieve the blind-spot property. Our network improves the performance over the prior work and achieves state-of-the-art results on established datasets.
We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in... more We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in thin gold films. The hole arrays are fabricated by a low-cost technique using convectively self-assembled polystyrene beads as an initial pattern template. Our results suggest that corrugated, short-range ordered hole arrays allow the excitation of resonant surface plasmons. This opens the door to inexpensive, large- scale and high throughput fabrication of nanophotonic devices.
Designing Deep Neural Networks (DNNs) running on edge hardware remains a challenge. Standard desi... more Designing Deep Neural Networks (DNNs) running on edge hardware remains a challenge. Standard designs have been adopted by the community to facilitate the deployment of Neural Network models. However, not much emphasis is put on adapting the network topology to fit hardware constraints. In this paper, we adapt one of the most widely used architectures for mobile hardware platforms, MobileNetV2, and study the impact of changing its topology and applying post-training quantization. We discuss the impact of the adaptations and the deployment of the model on an embedded hardware platform for face detection.
The Aeronautics industry has pioneered safety from digital checklists to moving maps that improve... more The Aeronautics industry has pioneered safety from digital checklists to moving maps that improve pilot situational awareness and support safe ground movements. Today, pilots deal with increasingly complex cockpit environments and air traffic densification. Here we present an intelligent vision system, which allows real-time human-machine interaction in the cockpits to reduce pilot’s workload. The challenges for such a vision system include extreme change in background light intensity, large field-of-view and variable working distances. Adapted hardware, use of state-of-the-art computer vision techniques and machine learning algorithms in eye gaze detection allow a smooth, and accurate real-time feedback system. The current system has been over-specified to explore the optimized solutions for different use-cases. The algorithmic pipeline for eye gaze tracking was developed and iteratively optimized to obtain the speed and accuracy required for the aviation use cases. The pipeline, which is a combination of data-driven and analytics approaches, runs in real time at 60 fps with a latency of about 32ms. The eye gaze estimation error was evaluated in terms of the point of regard distance error with respect to the 3D point location. An average error of less than 1.1cm was achieved over 28 gaze points representing the cockpit instruments placed at about 80-110cm from the participants’ eyes. The angular gaze deviation goes down to less than 1° for the panels towards which an accurate eye gaze was required according to the use cases.
We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in... more We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in thin gold films. The hole arrays are fabricated by a low-cost technique using convectively self-assembled polystyrene beads as an initial pattern template. Our results suggest that corrugated, short-range ordered hole arrays allow the excitation of resonant surface plasmons. This opens the door to inexpensive, largescale and high throughput fabrication of nanophotonic devices.
A photonic crystal structure for condensing electromagnetic radiation comprising a modulated squa... more A photonic crystal structure for condensing electromagnetic radiation comprising a modulated square-shaped lattice and a corresponding condensing device are disclosed
Multispectral imagers collect a hypercube of data, where the spatial image is along two-dimension... more Multispectral imagers collect a hypercube of data, where the spatial image is along two-dimensions and the spectral information is in the third. Two main technologies are used for multispectral imaging: sweeping, where the hypercube is built by scanning through different wavelengths or spatial positions and snapshot multispectral spectral imaging, where the 3D cube of images is taken in one shot. Sweeping imaging systems tend to have more lines and better spectral resolutions whilst snapshot cameras are often used for dynamic analysis of scenes. A common method to obtain the hypercube in snapshot imagers is by pixel level filtering on the sensor chip. Pixel level filtering, where the filter is placed directly on the pixels are intergrated into the wafer-level making processing making them difficult to customize. Therefore, these sensors tend to aim for equally spaced spectral lines in order to cover many applications. This results in an often in an unnecessarily large data cube when...
We have designed, fabricated and tested narrow-band Fabry-Perot filters in the infrared using gol... more We have designed, fabricated and tested narrow-band Fabry-Perot filters in the infrared using gold porous mirrors and a silicon spacer layer. The filter peaks at 10 μm and 15 μm have approximately 10% transmission and a 1.5% linewidth. A Fabry-Perot structure with plane metal layers having a similar linewidth would have a transmission of only 0.2%. Thus, for the same linewidth we have improved the transmission by a factor of 50. Apart from the optical enhancements, these filters also have the advantage that they can be made inexpensively in a standard silicon MEMS technology and that their resonances can be finely tuned through post processing.
In this paper, we experimentally studied both the bright-field and fluorescence images of microsp... more In this paper, we experimentally studied both the bright-field and fluorescence images of microspheres by conventional and confocal scanning polarization microscopes. A qualitative analysis have been given to show a physical picture on the imaging of the microspheres. Emission spectra from melamine formaldehyde microspheres stained with Ethidium Bromide or covered by thin shell of CdTe nanocrystals have been experimentally studied. We adopted analytical expressions describing the resonance spacing in order to determine the size of the microspheres.
Confocal, Multiphoton, and Nonlinear Microscopic Imaging II, 2005
Confocal microscopy has a unique optical sectioning property which allows three-dimensional image... more Confocal microscopy has a unique optical sectioning property which allows three-dimensional images at different depths. Use of a microlens array is a potential alternative to the Nipkow disk for parallel imaging with high throughput in real-time confocal microscopy. The use of variable-focal-length microlenses can provide a way to axially scan the foci electronically avoiding the inflexible mechanical movement of the
The use of variable-focal-length (VFL) microlenses can provide a way to axially scan the foci acr... more The use of variable-focal-length (VFL) microlenses can provide a way to axially scan the foci across a sample by electronic control. We demonstrate an approach to coupling VFL microlenses individually to a fiber bundle as a way to create a high-throughput aperture array with a controllable aperture pattern. It would potentially be applied in real-time confocal imaging in vivo for biological specimens. The VFL microlenses that we used consist of a liquid-crystal film sandwiched between a pair of conductive substrates for which one has a hole-patterned electrode. One obtains the variation of the focal length by changing the applied voltage. The fiber bundle has been characterized by coupling with both coherent and incoherent light sources. We further demonstrate the use of a VFL microlens array in combination with the fiber bundle to build up a confocal system. The axial response of the confocal system has been measured without mechanical movement of the sample or the objective, and the FWHM is estimated to be approximately 16 m, with asymmetric sidelobes.
2020 14th International Symposium on Medical Information Communication Technology (ISMICT), 2020
Monitoring of breathing rate considerably improves healthcare quality by providing information ab... more Monitoring of breathing rate considerably improves healthcare quality by providing information about the patient’s state. Yet, the current methods used to assess the respiratory rate have some drawbacks that could be improved. Recent developments of the non-contact based measurement of the respiratory rate brought the remote Photoplethysmography (rPPG) technique. This technique allows physiological parameters monitoring with a camera and ambient light. It measures the light absorption variation by the blood to extract several physiological parameters. The main issue with the rPPG method is that its signal quality is lower compared with the contact-based methods. Several ways to improve this quality may be explored to obtain a robust rPPG pipeline. One of these is the combination of the multidimensional temporal traces generated by a camera into a single temporal trace. The literature methods aimed to extract a signal with the maximal information using smart estimation of the combination coefficients. These coefficients are computed using statistical or physiological/optical properties of the traces. In this paper, we developed an algorithm that estimates a better respiratory signal than with regular methods. The algorithm named Energy Variance Maximization (EVM) estimates the linear combination that maximizes the Signal Noise Ratio (SNR) of the output trace. We compare our contribution with three state-of-the-art method: CHROM, PBV and PVM. The results obtained with this method are better than the state of the art methods, with the Mean Absolute Error (MAE) being 0.87 rpm better than CHROM, the best of compared literature methods.
The problem of estimating a surface shape from its observed reflectance properties still remains ... more The problem of estimating a surface shape from its observed reflectance properties still remains a challenging task in computer vision. The presence of global illumination effects such as inter-reflections or cast shadows makes the task particularly difficult for non-convex real-world surfaces. State-of-the-art methods for calibrated photometric stereo address these issues using convolutional neural networks (CNNs) that primarily aim to capture either the spatial context among adjacent pixels or the photometric one formed by illuminating a sample from adjacent directions. In this paper, we bridge these two objectives and introduce an efficient fully-convolutional architecture that can leverage both spatial and photometric context simultaneously. In contrast to existing approaches that rely on standard 2D CNNs and regress directly to surface normals, we argue that using separable 4D convolutions and regressing to 2D Gaussian heat-maps severely reduces the size of the network and make...
2020 7th Swiss Conference on Data Science (SDS), 2020
Image denoising is an essential tool in computational photography. Standard denoising techniques,... more Image denoising is an essential tool in computational photography. Standard denoising techniques, which use deep neural networks at their core, require pairs of clean and noisy images for its training. If we do not possess the clean samples, we can use blind-spot neural network architectures, which estimate the pixel value based on the neighbouring pixels only. These networks thus allow training on noisy images directly, as they by-design avoid trivial solutions. Nowadays, the blind-spot is mostly achieved using shifted convolutions or serialization. We propose a novel fully convolutional network architecture that uses dilations to achieve the blind-spot property. Our network improves the performance over the prior work and achieves state-of-the-art results on established datasets.
We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in... more We present the characterization of short-range ordered hexagonal arrays of subwavelength holes in thin gold films. The hole arrays are fabricated by a low-cost technique using convectively self-assembled polystyrene beads as an initial pattern template. Our results suggest that corrugated, short-range ordered hole arrays allow the excitation of resonant surface plasmons. This opens the door to inexpensive, large- scale and high throughput fabrication of nanophotonic devices.
Uploads
Papers by Andrea Dunbar