scholarly journals Encoding and Recall of Spatio-Temporal Episodic Memory in Real Time

2017 ◽  
Author(s):  
Poo-Hee Chang ◽  
Ah-Hwee Tan
Keyword(s):  
Episodic Memory ◽  
Real Time ◽  
State Of The Art ◽  
Contextual Information ◽  
Activity Patterns ◽  
Memory Model ◽  
Data Set ◽  
Episodic Events ◽  
Spatio Temporal ◽  
High Level

Episodic memory enables a cognitive system to improve its performance by reflecting upon past events. In this paper, we propose a computational model called STEM for encoding and recall of episodic events together with the associated contextual information in real time. Based on a class of self-organizing neural networks, STEM is designed to learn memory chunks or cognitive nodes, each encoding a set of co-occurring multi-modal activity patterns across multiple pattern channels. We present algorithms for recall of events based on partial and inexact input patterns. Our empirical results based on a public domain data set show that STEM displays a high level of efficiency and robustness in encoding and retrieval with both partial and noisy search cues when compared with a state-of-the-art associative memory model.

Spatio-Temporal Denoising for Depth Map Sequences

2016 ◽  
Vol 7 (2) ◽  
pp. 21-35
Author(s):  
Thomas Hach ◽  
Tamara Seybold
Keyword(s):  
State Of The Art ◽  
Synthetic Data ◽  
Depth Map ◽  
Data Set ◽  
Temporal Noise ◽  
Camera Systems ◽  
Spatio Temporal ◽  
Novel Strategy ◽  
Depth Video ◽  
High Level

This paper proposes a novel strategy for depth video denoising in RGBD camera systems. Depth map sequences obtained by state-of-the-art Time-of-Flight sensors suffer from high temporal noise. Hence, all high-level RGB video renderings based on the accompanied depth maps' 3D geometry like augmented reality applications will have severe temporal flickering artifacts. The authors approached this limitation by decoupling depth map upscaling from the temporal denoising step. Thereby, denoising is processed on raw pixels including uncorrelated pixel-wise noise distributions. The authors' denoising methodology utilizes joint sparse 3D transform-domain collaborative filtering. Therein, they extract RGB texture information to yield a more stable and accurate highly sparse 3D depth block representation for the consecutive shrinkage operation. They show the effectiveness of our method on real RGBD camera data and on a publicly available synthetic data set. The evaluation reveals that the authors' method is superior to state-of-the-art methods. Their method delivers flicker-free depth video streams for future applications.

scholarly journals Adaptive Multisensor Acquisition via Spatial Contextual Information for Compressive Spectral Image Classification

2021 ◽  
Author(s):  
Nelson Diaz ◽  
Juan Marcos ◽  
Esteban Vera ◽  
Henry Arguello
Keyword(s):  
Image Classification ◽  
State Of The Art ◽  
Contextual Information ◽  
Real Data ◽  
Experimental Setup ◽  
Spectral Image ◽  
Data Set ◽  
Adaptive Sensing

Results of extensive simulations are shown for two state-of-the-art databases: Pavia University and Indian Pines. Furthermore, an experimental setup that performs the adaptive sensing was built to test the performance of the proposed approach on a real data set.

scholarly journals Context-Aware Image Inpainting with Learned Semantic Priors

2021 ◽  
Author(s):  
Wendong Zhang ◽  
Junwei Zhu ◽  
Ying Tai ◽  
Yunbo Wang ◽  
Wenqing Chu ◽  
...  
Keyword(s):  
State Of The Art ◽  
Contextual Information ◽  
Image Inpainting ◽  
Context Aware ◽  
Global Context ◽  
Low Level ◽  
Complex Scenes ◽  
Knowledge Distillation ◽  
Image Generator ◽  
High Level

Recent advances in image inpainting have shown impressive results for generating plausible visual details on rather simple backgrounds. However, for complex scenes, it is still challenging to restore reasonable contents as the contextual information within the missing regions tends to be ambiguous. To tackle this problem, we introduce pretext tasks that are semantically meaningful to estimating the missing contents. In particular, we perform knowledge distillation on pretext models and adapt the features to image inpainting. The learned semantic priors ought to be partially invariant between the high-level pretext task and low-level image inpainting, which not only help to understand the global context but also provide structural guidance for the restoration of local textures. Based on the semantic priors, we further propose a context-aware image inpainting model, which adaptively integrates global semantics and local features in a unified image generator. The semantic learner and the image generator are trained in an end-to-end manner. We name the model SPL to highlight its ability to learn and leverage semantic priors. It achieves the state of the art on Places2, CelebA, and Paris StreetView datasets

scholarly journals Mid-IR spectrometer for mobile, real-time urban NO<sub>2</sub> measurements

2018 ◽  
Vol 11 (5) ◽  
pp. 2669-2681 ◽  
Author(s):  
P. Morten Hundt ◽  
Michael Müller ◽  
Markus Mangold ◽  
Béla Tuzson ◽  
Philipp Scheidegger ◽  
...  
Keyword(s):  
Air Quality ◽  
Urban Environment ◽  
Real Time ◽  
Infrared Laser ◽  
Detailed Knowledge ◽  
Data Set ◽  
Laser Absorption ◽  
Spatial Coverage ◽  
Spatio Temporal ◽  
The Impact

Abstract. Detailed knowledge about the urban NO2 concentration field is a key element for obtaining accurate pollution maps and individual exposure estimates. These are required for improving the understanding of the impact of ambient NO2 on human health and for related air quality measures. However, city-scale NO2 concentration maps with high spatio-temporal resolution are still lacking, mainly due to the difficulty of accurate measurement of NO2 at the required sub-ppb level precision. We contribute to close this gap through the development of a compact instrument based on mid-infrared laser absorption spectroscopy. Leveraging recent advances in infrared laser and detection technology and a novel circular absorption cell, we demonstrate the feasibility and robustness of this technique for demanding mobile applications. A fully autonomous quantum cascade laser absorption spectrometer (QCLAS) has been successfully deployed on a tram, performing long-term and real-time concentration measurements of NO2 in the city of Zurich (Switzerland). For ambient NO2 concentrations, the instrument demonstrated a precision of 0.23 ppb at one second time resolution and of 0.03 ppb after 200 s averaging. Whilst the combined uncertainty estimated for the retrieved spectroscopic values was less than 5 %, laboratory intercomparison measurements with standard CLD instruments revealed a systematic NO2 wall loss of about 10 % within the laser spectrometer. For the field campaign, the QCLAS has been referenced to a CLD using urban atmospheric air, despite the potential cross sensitivity of CLD to other nitrogen containing compounds. However, this approach allowed a direct comparison and continuous validation of the spectroscopic data to measurements at regulatory air quality monitoring (AQM) stations along the tram-line. The analysis of the recorded high-resolution time series allowed us to gain more detailed insights into the spatio-temporal concentration distribution of NO2 in an urban environment. Furthermore, our results demonstrate that for reliable city-scale concentration maps a larger data set and better spatial coverage is needed, e.g., by deploying more mobile and stationary instruments to account for mainly two shortcomings of the current approach: (i) limited residence time close to sources with large short-term NO2 variations, and (ii) insufficient representativeness of the tram tracks for the complex urban environment.

Real-Time Measurement of Tourists’ Objective and Subjective Emotions in Time and Space

2017 ◽  
Vol 57 (1) ◽  
pp. 3-16 ◽  
Author(s):  
Noam Shoval ◽  
Yonatan Schvimer ◽  
Maya Tamir
Keyword(s):  
Real Time ◽  
Electrodermal Activity ◽  
Contextual Information ◽  
Physiological Measures ◽  
Time And Space ◽  
Tourist Attractions ◽  
Individual Level ◽  
Real Time Measurement ◽  
Spatio Temporal ◽  
The Individual

The examination of tourists’ experiences is an essential subject in tourism scholarship. This study presents novel methods by which spatio-temporal data can be combined with physiological measures of emotion and semantic contextual information in order to obtain a comprehensive and integrative understanding of tourists’ experience in time and space. Four data collection techniques were combined and applied to a sample of 68 tourists in Jerusalem: high-resolution locational data, real-time surveying techniques using the experience sampling method, physiological measures of emotion (electrodermal activity), and traditional surveying techniques. We present methods for using these techniques in exploring data on the individual level, comparing pairs of individuals, and examining a sample, providing insight both on the individual’s personal experience and, more broadly, on the emotional characteristics of locations and tourist attractions in a city. Theoretical and methodological implications as well as the limitations of these techniques are discussed.

A Hierarchical Sequence-to-Sequence Model for Korean POS Tagging

2021 ◽  
Vol 20 (2) ◽  
pp. 1-13
Author(s):  
Guozhe Jin ◽  
Zhezhou Yu
Keyword(s):  
Language Processing ◽  
Morphological Analysis ◽  
State Of The Art ◽  
Contextual Information ◽  
Surface Form ◽  
Pos Tagging ◽  
Part Of Speech ◽  
Base Form ◽  
High Level ◽  
Tagging Methods

Part-of-speech (POS) tagging is a fundamental task in natural language processing. Korean POS tagging consists of two subtasks: morphological analysis and POS tagging. In recent years, scholars have tended to use the seq2seq model to solve this problem. The full context of a sentence is considered in these seq2seq-based Korean POS tagging methods. However, Korean morphological analysis relies more on local contextual information, and in many cases, there exists one-to-one matching between morpheme surface form and base form. To make better use of these characteristics, we propose a hierarchical seq2seq model. In our model, the low-level Bi-LSTM encodes the syllable sequence, whereas the high-level Bi-LSTM models the context information of the whole sentence, and the decoder generates the morpheme base form syllables as well as the POS tags. To improve the accuracy of the morpheme base form recovery, we introduced the convolution layer and the attention mechanism to our model. The experimental results on the Sejong corpus show that our model outperforms strong baseline systems in both morpheme-level F1-score and eojeol-level accuracy, achieving state-of-the-art performance.

scholarly journals Adaptive Multisensor Acquisition via Spatial Contextual Information for Compressive Spectral Image Classification

2021 ◽  
Author(s):  
Nelson Diaz ◽  
Juan Marcos ◽  
Esteban Vera ◽  
Henry Arguello
Keyword(s):  
Image Classification ◽  
State Of The Art ◽  
Contextual Information ◽  
Real Data ◽  
Experimental Setup ◽  
Spectral Image ◽  
Data Set ◽  
Adaptive Sensing

Results of extensive simulations are shown for two state-of-the-art databases: Pavia University and Indian Pines. Furthermore, an experimental setup that performs the adaptive sensing was built to test the performance of the proposed approach on a real data set.

scholarly journals Real time rainfall estimation using microwave signals of cellular communication networks: a case study of Faisalabad, Pakistan

2018 ◽  
Author(s):  
Muhammad Sohail Afzal ◽  
Syed Hamid Hussain Shah ◽  
Muhammad Jehanzeb Masud Cheema ◽  
Riaz Ahmad
Keyword(s):  
Water Balance ◽  
Real Time ◽  
Rainfall Intensity ◽  
Rain Gauge ◽  
Cellular Communication ◽  
Rainfall Estimation ◽  
Data Set ◽  
Spatial Error ◽  
Near Surface ◽  
Spatio Temporal

Abstract. Water balance estimate requires high spatio-temporal water balance components and rainfall is one of them. Rainfall is stochastic variable, which varies with respect to space and time. There are different methods for rainfall estimation such as rain gauge, satellite data but the resolution of these methods are very low, which cause over and underestimation of rainfall. A real time rainfall estimation mechanism is tested using commercial cellular networks in Faisalabad, district of Pakistan. The microwave links are used to quantify rainfall intensities and estimate rainfall at high spatio-temporal resolution. The attenuation in electromagnetic signals due to varying rainfall intensities is measured by taking difference between the power transmitted and power received during rainy period and is the measure of the path-averaged rainfall intensity. This rainfall related distortion is converted into rainfall intensity. This technique is applied on a standard microwave communication network used by a cellular communication system, comprising 35 microwave links, and it allow for observation of near-surface rainfall at the temporal resolutions of 15 min. Signal data-set of year 2012–2014 and 2015–2017 is used for calibration and validation respectively with three rain gauge data-set. The accuracy of the method is demonstrated by comparing the daily cumulative rainfall depth of University of Agriculture Faisalabad rain gauge (UAF-RG), Ayub Agriculture Research rain gauge(AR-RG) and Water and Sanitation Authority rain gauge (WASA-RG) with link based rainfall depths estimated from L2, L28 and L34 respectively, reaching r2 up to 0.97. UAF-RG is considered reference to study the spatial variability of rainfall of all the selected links within the study area, observed 10 %–60 % average spatial error of all links with the reference UAF-RG. All the results show that microwave links are potentially useful compared to the low resolution methods of rainfall estimation and can be used for effective water resources management.

scholarly journals Reconstructing feedback representations in the ventral visual pathway with a generative adversarial autoencoder

2021 ◽  
Vol 17 (3) ◽  
pp. e1008775
Author(s):  
Haider Al-Tahan ◽  
Yalda Mohsenzadeh
Keyword(s):  
Visual Information ◽  
Activity Patterns ◽  
Visual Pathway ◽  
Generative Model ◽  
Data Set ◽  
Low Level ◽  
Ventral Visual Pathway ◽  
Neural Feedback ◽  
High Level ◽  
Diverse Data

While vision evokes a dense network of feedforward and feedback neural processes in the brain, visual processes are primarily modeled with feedforward hierarchical neural networks, leaving the computational role of feedback processes poorly understood. Here, we developed a generative autoencoder neural network model and adversarially trained it on a categorically diverse data set of images. We hypothesized that the feedback processes in the ventral visual pathway can be represented by reconstruction of the visual information performed by the generative model. We compared representational similarity of the activity patterns in the proposed model with temporal (magnetoencephalography) and spatial (functional magnetic resonance imaging) visual brain responses. The proposed generative model identified two segregated neural dynamics in the visual brain. A temporal hierarchy of processes transforming low level visual information into high level semantics in the feedforward sweep, and a temporally later dynamics of inverse processes reconstructing low level visual information from a high level latent representation in the feedback sweep. Our results append to previous studies on neural feedback processes by presenting a new insight into the algorithmic function and the information carried by the feedback processes in the ventral visual pathway.

Contextual location prediction using spatio-temporal clustering

2016 ◽  
Vol 12 (3) ◽  
pp. 290-309 ◽  
Author(s):  
Djamel Guessoum ◽  
Moeiz Miraoui ◽  
Chakib Tadj
Keyword(s):  
Pervasive Computing ◽  
Spatial Clustering ◽  
Contextual Information ◽  
Real Data ◽  
Temporal Segmentation ◽  
Data Set ◽  
Content Type ◽  
Points Of Interest ◽  
Day Of The Week ◽  
Spatio Temporal

Purpose The prediction of a context, especially of a user’s location, is a fundamental task in the field of pervasive computing. Such predictions open up a new and rich field of proactive adaptation for context-aware applications. This study/paper aims to propose a methodology that predicts a user’s location on the basis of a user’s mobility history. Design/methodology/approach Contextual information is used to find the points of interest that a user visits frequently and to determine the sequence of these visits with the aid of spatial clustering, temporal segmentation and speed filtering. Findings The proposed method was tested with a real data set using several supervised classification algorithms, which yielded very interesting results. Originality/value The method uses contextual information (current position, day of the week, time and speed) that can be acquired easily and accurately with the help of common sensors such as GPS.

Sign in / Sign up

Export Citation Format

Share Document