From the Publisher: 
This is undoubtedly the most accessible book on digital signal processing (DSP) available to the beginner. Using intuitive explanations and well-chosen examples, this book gives you the tools to develop a fundamental understanding of DSP theory. The author covers the essential mathematics by explaining the meaning and significance of the key DSP equations. Comprehensive in scope, and gentle in approach, the book will help you achieve a thorough grasp of the basics and move gradually to more sophisticated DSP concepts and applications.

数字信号处理 = Understanding digital signal processing

The Tell-Tale Brain by V.S. Ramachandran delivers a thought-provoking exploration of the neural basis of human cognition. Ramachandran, a prominent cognitive neuroscientist and practicing neurologist at UCSD, is well known for his studies on phantom limb, in which sensations seem to arrive from amputated body parts, and synesthesia, in which sounds may seem to have tastes, letters may have colors, or there may be some other combinations of actual and illusory events. These and other neurological oddities serve as a jumping off point for the book’s examination of the human brain, how it evolved, and what mechanisms might differentiate us from “lower” animals. 

Writing for a general audience as well as neuroscience specialists, Ramachandran attempts to go beyond what is known and speculate about what might be. He argues, “If every scientist were a brush-clearing, never-stray-beyond-established-fact type, science would advance at a snail’s pace and would have a hard time unpainting itself out of corners.” In an era in which shrinking availability of research funding induces greater and greater conservatism, Ramachandran’s open-minded approach provides a refreshing contrast. This book expands the realm of ideas and topics that seem tractable to scientific exploration. That this may be accompanied by some overreaching comes with the territory; indeed, in a work of this kind, failing to overreach would be the greater error.

The book opens with the challenge of understanding what it is about our “three pound mass of jelly” that affords us our dazzling array of cognitive abilities. The accelerating pace of advancement over the last few thousand years, Ramachandran argues, requires some explanation. During this period, we have perfected cultural transmission of information, so that individuals can benefit from the hard-earned knowledge of others. What is it about our brains that allowed this to happen? Ramachandran points to two attributes of the brain as notable contributors.

The first is that neural connections don’t just go where they are supposed to. As both phantom limb syndrome and synesthesia demonstrate, some of the connections in the brain are in the “wrong” place. The limb region of the somatosensory cortex also gets a whisper of input from the face, a pathway only revealed when the limb is missing. Synesthetes experience perceptual cross-talk between real stimuli and illusory, seemingly unrelated sensations. Ramachandran suggests that such anomalous connections reveal a propensity for forging ties between content of different kinds, a propensity that would serve us well in the development of new cognitive abilities. The human brain might have a kind of anatomical flexibility that could lead, on an evolutionary and perhaps on an individual time scale, to repurposing brain areas to perform new tasks. If the connections are already there, evolution needn’t build a whole new infrastructure for some novel cognitive function from scratch.

The second major contributors to the evolution of human cognition, Ramachandran argues, are specialized neurons known as “mirror” neurons. These neurons were first discovered in the ventral premotor cortex in monkeys in 1992 by an Italian research group led by Giaocomo Rizzolatti. Mirror neurons show increased activity both when the monkey performs an action and when it watches someone else perform that same action. Ramachandran suggests that by lumping observation and performance, mirror neurons might provide us the basis for empathy, theory of mind, and the ability to predict the actions and motivations of others — vital elements of primate, and human, social structure.

The last few chapters of the book shift focus to explore the neuroscience of art — why do we like to look at what we like to look at? Here, Ramachandran proposes that what we find pleasing follows some rules, similar to the Gestalt grouping principles. Our visual systems “like” to detect objects, to see contrast, to solve “peekaboo” problems, etc. One of the most intriguing ideas in this section is that our aesthetic preferences might be driven by canonical ideals representing extreme forms of real-world stimuli. For example, in the 1940s, Dutch ethologist Niko Tinbergen found that seagull chicks pecked at stimuli resembling the red dot on their parent’s bill, but they pecked even harder at an abstract representation of the bill. Ramachandran argues that a similar tendency could explain our enjoyment of caricatures and other exaggerated artistic forms. This idea suggests a path forward for investigating the neural basis of our aesthetic preferences — currently, very much an unsolved problem — by relating aesthetics to the tuning properties of neurons in the visual pathway.

Not surprisingly, some arguments in this book have a better foundation than others (a section on autism is likely to be particularly controversial), but in the end, the most important contribution of this creative book lies not just in the particular ideas that Ramachandran offers, but in the ideas that he provokes his readers to come up with. Looking at things in a new way may lead us to unpaint a few corners, and, perhaps eventually, to clear brush in some new domains of neuroscience.

/pdf/the-tell-tale-brain-a-neuroscientist-s-quest-for-what-makes-wdoodds3np.pdf

The tell-tale brain: A neuroscientist’s quest for what makes us human

The 2004 ACM SIGMOD International Conference on Management of Data, held in Paris in the week of June 13-18, is the first SIGMOD ever held outside of North America, and it has chosen a place that is rich in tradition but also rich in new departures, one of the focal points of the age of enlightenment and the place of the French revolution in 1789. We hope that this is a symbol of SIGMOD embracing new challenges in the broader area of business and scientific information management and the wider spectrum of network-enabled data-rich applications.The SIGMOD conference is sponsored by the Association for Computing Machinery (ACM) and its Special Interest Group on Management of Data (SIGMOD). Two days of the conference overlap with the Symposium on Principles of Database System (PODS), and several workshops and other events are co-located with SIGMOD.

Proceedings of the 2004 ACM SIGMOD international conference on Management of data

This paper reports on Task 1 of the 2017 CLEF eHealth evaluation lab which extended the previous information extraction tasks of ShARe/CLEF eHealth evaluation labs. The task continued with codingof death certificates, as introduced in CLEF eHealth 2016. This large-scale classification task consisted of extracting causes of death as coded in the International Classification of Diseases, tenth revision (ICD10). The languages offered for the task this year were English and French. Participant systems were evaluated against a blind reference standardof 31,690 death certificates in the French dataset and 6,665 certificates in the English dataset using Precision, Recall and F-measure. In total, eleven teams participated: 10 teams submitted runs for the English dataset and 9 for the French dataset. Five teams submitted their systemsto the reproducibility track. For death certificate coding, the highest performance was 0.8674 F-measure for French and 0.8501 for English.

/pdf/clef-ehealth-2017-multilingual-information-extraction-task-3zc8m05nkr.pdf

CLEF eHealth 2017 Multilingual Information Extraction task overview: ICD10 coding of death certificates in English and French

This study had three purposes, all related to evaluating energy expenditure (EE) prediction accuracy from body-worn accelerometers: (1) compare linear regression to linear mixed models, (2) compare linear models to artificial neural network models, and (3) compare accuracy of accelerometers placed on the hip, thigh, and wrists. Forty individuals performed 13 activities in a 90 min semi-structured, laboratory-based protocol. Participants wore accelerometers on the right hip, right thigh, and both wrists and a portable metabolic analyzer (EE criterion). Four EE prediction models were developed for each accelerometer: linear regression, linear mixed, and two ANN models. EE prediction accuracy was assessed using correlations, root mean square error (RMSE), and bias and was compared across models and accelerometers using repeated-measures analysis of variance. For all accelerometer placements, there were no significant differences for correlations or RMSE between linear regression and linear mixed models (correlations: r  =  0.71-0.88, RMSE: 1.11-1.61 METs; p  >  0.05). For the thigh-worn accelerometer, there were no differences in correlations or RMSE between linear and ANN models (ANN-correlations: r  =  0.89, RMSE: 1.07-1.08 METs. Linear models-correlations: r  =  0.88, RMSE: 1.10-1.11 METs; p  >  0.05). Conversely, one ANN had higher correlations and lower RMSE than both linear models for the hip (ANN-correlation: r  =  0.88, RMSE: 1.12 METs. Linear models-correlations: r  =  0.86, RMSE: 1.18-1.19 METs; p  <  0.05), and both ANNs had higher correlations and lower RMSE than both linear models for the wrist-worn accelerometers (ANN-correlations: r  =  0.82-0.84, RMSE: 1.26-1.32 METs. Linear models-correlations: r  =  0.71-0.73, RMSE: 1.55-1.61 METs; p  <  0.01). For studies using wrist-worn accelerometers, machine learning models offer a significant improvement in EE prediction accuracy over linear models. Conversely, linear models showed similar EE prediction accuracy to machine learning models for hip- and thigh-worn accelerometers and may be viable alternative modeling techniques for EE prediction for hip- or thigh-worn accelerometers.

https://iopscience.iop.org/article/10.1088/1361-6579/38/2/343/pdf

Comparison of linear and non-linear models for predicting energy expenditure from raw accelerometer data

Periodicities (repeating patterns) are observed in many human behaviors. Their strength may capture untapped patterns that incorporate sleep, sedentary, and active behaviors into a single metric indicative of better health. We present a framework to detect periodicities from longitudinal wrist-worn accelerometry data. GENEActiv accelerometer data were collected from 20 participants (17 men, 3 women, aged 35-65) continuously for 64.4 ± 26.2 (range: 13.9 to 102.0) consecutive days. Cardiometabolic risk biomarkers and health-related quality of life metrics were assessed at baseline. Periodograms were constructed to determine patterns emergent from the accelerometer data. Periodicity strength was calculated using circular autocorrelations for time-lagged windows. The most notable periodicity was at 24 h, indicating a circadian rest-activity cycle; however, its strength varied significantly across participants. Periodicity strength was most consistently associated with LDL-cholesterol (r's = 0.40-0.79, P's < 0.05) and triglycerides (r's = 0.68-0.86, P's < 0.05) but also associated with hs-CRP and health-related quality of life, even after adjusting for demographics and self-rated physical activity and insomnia symptoms. Our framework demonstrates a new method for characterizing behavior patterns longitudinally which captures relationships between 24 h accelerometry data and health outcomes.

/pdf/behavioral-periodicity-detection-from-24-h-wrist-1ng9ww5qvt.pdf

Behavioral Periodicity Detection from 24 h Wrist Accelerometry and Associations with Cardiometabolic Risk and Health-Related Quality of Life

Memorability, defined as the quality of being worth remembering, is a pressing issue in media as we struggle to organize and retrieve digital content and make it more useful in our daily lives. The Predicting Media Memorability task in MediaEval 2019 tackles this problem by creating a challenge to automatically predict memorability scores building on the work developed in 2018. Our team ensembled transfer learning approaches with video captions using embeddings and our own pre-computed features which outperformed Medieval 2018’s state-of-the-art architectures.

/pdf/predicting-media-memorability-using-ensemble-models-32xknsogh9.pdf

Predicting Media Memorability Using Ensemble Models.

Lifelogging is the ambient, continuous digital recording of a person's everyday activities for a variety of possible applications. Much of the work to date in lifelogging has focused on developing sensors, capturing information, processing it into events and then supporting event-based access to the lifelog for applications like memory recall, behaviour analysis or similar. With the recent arrival of aggregating platforms such as Apple's HealthKit, Microsoft's HealthVault and Google's Fit, we are now able to collect and aggregate data from lifelog sensors, to centralize the management of data and in particular to search for and detect patterns of usage for individuals and across populations. In this paper, we present a framework that detects both low-level and high-level periodicity in lifelog data, detecting hidden patterns of which users would not otherwise be aware. We detect periodicities of time series using a combination of correlograms and periodograms, using various signal processing algorithms. Periodicity detection in lifelogs is particularly challenging because the lifelog data itself is not always continuous and can have gaps as users may use their lifelog devices intermittingly. To illustrate that periodicity can be detected from such data, we apply periodicity detection on three lifelog datasets with varying levels of completeness and accuracy.

/pdf/periodicity-detection-in-lifelog-data-with-missing-and-111zpjojef.pdf

Periodicity detection in lifelog data with missing and irregularly sampled data

Sensor technologies can enable independent living for people with dementia by monitoring their behaviour and identifying points where support may be required. Wearable sensors can provide such support but may constitute a source of stigma for the user if they are perceived as visible and therefore obtrusive. This paper presents an initial empirical investigation exploring the extent to which wearable sensors are perceived as visible. 23 Participants wore eye tracking glasses, which superimposed the location of their gaze onto video data of their panorama. Participants were led to believe that the research entailed a subjective evaluation of the eye tracking glasses. A researcher wore one of two wearable sensors during the evaluation enabling us to measure the extent to which participants fixated on the sensor during a one-on-one meeting. Results are presented on the general visibility and potential fixations on two wearable sensors, a wrist-work actigraph and a lifelogging camera, during normal conversation between two people. Further investigation is merited according to the results of this pilot study.

/pdf/visibility-of-wearable-sensors-as-measured-using-eye-5bl9xbmrxo.pdf

Visibility of wearable sensors as measured using eye tracking glasses

Visual lifelogging using wearable cameras accumulates large amounts of image data. To make them useful they are typically structured into events corresponding to episodes which occur during the wearer’s day. These events can be represented as a visual storyboard, a collection of chronologically ordered images which summarise the day’s happenings. In previous work, little attention has been paid to how to select the representative keyframes for a lifelogged event, apart from the fact that the image should be of good quality in terms of absence of blurring, motion artifacts, etc. In this paper we look at image aesthetics as a characteristic of wearable camera images. We show how this can be used in combination with content analysis and temporal offsets, to offer new ways for automatically selecting wearable camera keyframes. In this paper we implement several variations of the keyframe selection method and illustrate how it works using a publicly-available lifelog dataset.

/pdf/image-aesthetics-and-content-in-selecting-memorable-3mzvx64ch6.pdf

Feiyan Hu

Papers

Behavioral Periodicity Detection from 24 h Wrist Accelerometry and Associations with Cardiometabolic Risk and Health-Related Quality of Life

Predicting Media Memorability Using Ensemble Models.

Periodicity detection in lifelog data with missing and irregularly sampled data

Visibility of wearable sensors as measured using eye tracking glasses

Image aesthetics and content in selecting memorable keyframes from lifelogs