Lino Ferreira

A generic framework for optimal 2D/3D key-frame extraction driven by aggregated saliency maps

This paper proposes a generic framework for extraction of key-frames from 2D or 3D video sequences, relying on a new method to compute 3D visual saliency. The framework comprises the following novel aspects that distinguish this work from previous ones: (i) the key-frame selection process is driven by an aggregated saliency map, computed from various feature maps, which in turn correspond to different visual attention models; (ii) a method for computing aggregated saliency maps in 3D video is proposed and validated using fixation density maps, obtained from ground-truth eye-tracking data; (iii) 3D video content is processed within the same framework as 2D video, by including a depth feature map into the aggregated saliency. A dynamic programming optimisation algorithm is used to find the best set of K frames that minimises the dissimilarity error (i.e., maximise similarity) between the original video shots of size N K and those reconstructed from the key-frames. Using different performance metrics and publicly available databases, the simulation results demonstrate that the proposed framework outperforms similar state-of-art methods and achieves comparable performance as other quite different approaches. Overall, the proposed framework is validated for a wide range of visual content and has the advantage of being independent from any specific visual saliency model or similarity metrics. HighlightsA 2D/3D key-frame extraction framework driven by aggregated saliency maps.A method to compute aggregated saliency maps in 3D video using attention models.Optimal key-frame extraction taking into account different visual saliency regions.

Efficient scalable coding of video summaries using dynamic GOP structures

A method to efficiently encode an arbitrary video summary with temporal scalability and dynamic Group of Pictures (GOP) structures is proposed in this paper. The video summary is encoded as the base layer of a Scalable Video Coding (SVC) bitstream using a novel approach, which matches the summary frames available in temporal segments onto corresponding dynamic GOP structures. An algorithm is devised to compute variable GOP sizes along with an efficient independent prediction structure for the summary. The results show that the proposed method can be used to encode arbitrary video summaries with increased efficiency in the temporal base layer and negligible loss of R-D performance in the whole scalable sequence.

Fibre optic hot-wire flowmeter based on a metallic coated hybrid LPG-FBG structure

In this work an all optical hot-wire flowmeter based on a silver coated fibre incorporating a long period grating and a Bragg grating is demonstrated. Optical energy at 1480 nm propagating down the fibre is coupled by the long period grating into the fibre cladding and absorbed by the metallic coating deposited on the fibre surface over the Bragg grating position. This absorption acts like a hot-wire raising locally the fibre temperature, which is effectively detected by the FBG resonance shift. The temperature raise depends on the flow speed of the surrounding air that has the effect to cool the fibre. In this way, the FBG Bragg wavelength shift can be related with the flow speed. Results obtained demonstrate the working principle and a flow speed resolution of 0.08 m/s is demonstrated.

Optimal priority MDC video streaming for networks with path diversity

This paper proposes a robust video streaming scheme for priority networks with path diversity, based on a combined approach of multiple description coding (MDC) with optimal picture classification into two priorities. A binary classification algorithm is proposed to define high (HP) and low (LP) priority network abstraction layer units (NALU), which in turn define the packet priorities. An optimisation algorithm is used to find HP pictures, based on dynamic programming and relying on minimisation of the packet loss concealment distortion. The paper shows that the proposed algorithm is able to effectively improve the decoded video without increasing the MDC stream redundancy. The overall performance evaluation, carried out by simulating MDC video streaming over lossy networks with path diversity, demonstrates that the proposed algorithm yields higher video quality for a wide range of packet loss rates (PLR). Comparing with no-priority MDC video streaming schemes, the simulation results show that the proposed algorithm can improve the average PSNR results in 0.7-3.2dB for packet loss rates between 3% and 15%.

Temperature- and strain-independent curvature sensor based on multimode interference

This work describes a fibre optic sensing structure that is sensitive to curvature, and features a low temperature- and strain cross-sensitivity. It is based on multimode interference, and relies on a single mode -- step index multimode -- single mode fibre configuration. It was observed that the transmitted optical power in such layout becomes highly sensitive to the wavelength of operation and to the length of the multimode fibre. The optical spectrum entertains two dominant loss bands, at wavelengths that have similar responses both to temperature and strain, and different responses to curvature. Based on this result, an interrogation approach is proposed that permits substantial sensitivity to curvature (8.7±0.1 nm.m) and residual sensitivities to temperature and strain (0.3±0.1 pm/°C and -0.06±0.01 pm/με respectively).

Fibre laser sensor based on a phase-shifted chirped grating for acoustic sensing of partial discharges in power transformers

Partial discharges are a major source of insulation failure in electric power transformers. Their detection can be achieved through the associated acoustic emissions, and this work reports on the investigation of a fibre laser sensor based on a phase-shifted chirped fibre grating for acoustic emission detection in the power transformer environment. The performance of the sensing head is characterized and compared for different surrounding media: air, water and oil.

A method for automatic detection of rectangular regions of interest in arbitrary images

This paper presents a computational method to extract optimum rectangular Regions of Interest (RoI) in images with an associated saliency map. Although saliency maps provide an individual relevance measure for each pixel, to find the sub-image (i.e., rectangular region) that contains the set of the most relevant pixels requires an optimisation procedure to define the boundaries of the best RoI. This is achieved by the method devised in the paper, by following an approach based on balancing the amount of relevant information that is included and excluded from the RoI. The results show that such method is capable of finding the most relevant rectangular RoI and thus to extract the optimum sub-images according to the relevance measure given by a generic saliency map. Since the method is not tied to any particular type of images, it finds application in quite different fields, such as salient object extraction and processing in industry and surveillance, image compression using attention modelling, biomedical imaging, etc.This paper presents a computational method to extract optimum rectangular Regions of Interest (RoI) in images with an associated saliency map. Although saliency maps provide an individual relevance measure for each pixel, to find the sub-image (i.e., rectangular region) that contains the set of the most relevant pixels requires an optimisation procedure to define the boundaries of the best RoI. This is achieved by the method devised in the paper, by following an approach based on balancing the amount of relevant information that is included and excluded from the RoI. The results show that such method is capable of finding the most relevant rectangular RoI and thus to extract the optimum sub-images according to the relevance measure given by a generic saliency map. Since the method is not tied to any particular type of images, it finds application in quite different fields, such as salient object extraction and processing in industry and surveillance, image compression using attention modelling, biomedical...

3D key-frame extraction method based on visual saliency

This paper presents a method for key-frame extraction from 3D video using visual saliency to weight the 3D content according to a user attention model. Key-frames are found in temporal segments of arbitrary length (i.e., 3D scenes) using a dynamic programming algorithm which minimises the dissimilarity between the reconstructed and the original temporal segment. The dissimilarity measure is based on a combination of frame difference and visual relevance estimated through visual saliency maps. These maps result from attention modeling, taking into account spatial, temporal and depth features of the 3D video content. The results, evaluated using the Shot Reconstruction Degree and the Fidelity measure, show that the proposed method outperforms those obtained from uniform sampling and attention curve methods. This method may be useful for fast browsing of 3D video repositories.

All-fibre interferometric configurations based on suspended-core fibres for pressure measurement

In this work, two all-fibre interferometric configurations based on suspended core fibres (SCF) are investigated. A Fabry-Pérot cavity made of SCF spliced in-between segments of single-mode and hollow-core fiber is proposed. An alternative sensing head configuration formed by the insertion of a length of SCF as a birefringence element in a Sagnac interferometer is also demonstrated. The sensitivity to pressure and temperature was determined for both configurations.

Fibre optic remote sensing based on long period gratings with in situ optical source

In this work the concept of long period based optical fibre sensors with the broadband light illumination generated just after the sensing structure is presented. This new approach allows the interrogation in transmission of the sensing head while integrated in a reflective configuration, which means the LPG sensor is seen in transmission by the optical source but in reflection by the measurement system. Also, it is shown that with this illumination layout the optical power balance is more favorable when compared with the standard configurations, allowing better sensor performances particularly when the sensing head is located far away from the photodetection and processing unit. This is demonstrated for the case of the LPG structure applied to measure strain and using ratiometric interrogation based on the readout of the optical power reflected by two fibre Bragg gratings spectrally located in each side of the LPG resonance.