Stairway to Elders: Bridging Space, Time and Emotions in Their Social

Environment for Wellbeing

Giuseppe Boccignone

1 a

, Claudio de’Sperati

2 b

, Marco Granato

1 c

, Giuliano Grossi

1 d

Raffaella Lanzarotti

1 e

, Nicoletta Noceti

3 f

and Francesca Odone

3 g

Department of Computer Science, Universit

a degli Studi di Milano, Milan, Italy

Department of Psychology, Universit

a Vita-Salute San Raffaele, Milan, Italy

MaLGa Center - DIBRIS - Universit

a di Genova, Genova, Italy

Keywords:

Emotion Recognition, Social Attitude, Video Speed Tuning, Emotion Regulation, Emotion-aware Ambient

Intelligence (AmE), Multi-modal Data, Gerontechnology.

Abstract:

The physical and mental health in elderly population is an emergent issue which in recent years has become

an urgent socio-economic phenomenon. Computer scientists, together with physicians and caregivers have

devoted a great research effort to conceive and devise assistive technologies, aiming at safeguarding elder

health, while a marginal consideration has been devoted to their emotional domain. In this manuscript we

outline the research plan and the objectives of a current project called Stairway to elders: bridging space, time

and emotions in their social environment for wellbeing”. Through a set of sensors, which include cameras and

physiological sensors, we aim at developing computational methods for understanding the affective state and

socialization attitude of older people in ecological conditions. A valuable by-product of the project will be the

collection of a multi-modal dataset to be used for model design, and that will be made available to the research

community. The outcomes of the project should support the design of an environment which automatically (or

semi-automatically) adapts its conditions to the affective state of older people, with a consequent improvement

of their life quality.

1 INTRODUCTION

Nowadays, especially in developed countries, life ex-

pectancy keeps growing, augmenting the proportion

of older people over the population. According to the

WHO, the worlds elderly population (deﬁned as peo-

ple aged 60 and older) has increased drastically in the

past decades and will reach about 2 billion in 2050. In

Europe, the percentage of the EU27 population above

65 years of age is foreseen to rise to 30% by 2060 (Gi-

annakouris et al., 2008). This evidence has opened a

social debate about how to face this socio-economic

phenomenon.

https://orcid.org/0000-0002-5572-0924

https://orcid.org/0000-0002-7322-2240

https://orcid.org/0000-0002-7322-4350

https://orcid.org/0000-0001-9274-4047

https://orcid.org/0000-0002-8534-4413

https://orcid.org/0000-0002-6482-4768

https://orcid.org/0000-0002-3463-2263

Indeed, the human being is a complex organism,

whose wellbeing may be described following several

dimensions, encompassing the physical, psychologi-

cal, economic, and social domains. The process of ag-

ing typically reduces the individuals potential in one

or more of these domains, leading to a condition of

vulnerability and clinical instability, deﬁned as frailty

in the specialized literature (Fried et al., 2001). In the

last decades, a great research effort has been devoted

to the design of assistive technologies that have pos-

itive impacts on different dimensions of health and

quality of life for aging population (Yared and Ab-

dulrazak, 2016). Also, the concept of smart envi-

ronments design has opened the possibility of mon-

itoring patients at home (Scanaill et al., 2006). To

date, a vast majority of methods proposed in this ﬁeld

addresses the problem of monitoring health status of

people mainly considering physical attributes while

the emotional and social domains have received only

marginal consideration.

Emotional/affective wellbeing is deeply associ-

548

Boccignone, G., de’Sperati, C., Granato, M., Grossi, G., Lanzarotti, R., Noceti, N. and Odone, F.

Stairway to Elders: Bridging Space, Time and Emotions in Their Social Environment for Wellbeing.

DOI: 10.5220/0009106605480554

In Proceedings of the 9th International Conference on Pattern Recognition Applications and Methods (ICPRAM 2020), pages 548-554

ISBN: 978-989-758-397-1; ISSN: 2184-4313

ated with the sense of fulﬁlment, including satisfac-

tion, optimism, having a purpose in life as well as be-

ing able to make the most of your abilities to cope

with the normal challenges of life. As suggested in

(Hawkins, 2005), the mental wellbeing of the aging

population is, together with health, a key factor for ag-

ing well since, besides irrefutable physical needs, the

wellbeing of a person cannot leave out emotional and

social aspects (Anttonen and Surakka, 2007). Unfor-

tunately, psychological stability is often undermined

over the years, when older people have to face life

challenges that weaken their independence and self-

conﬁdence.

These considerations are at the core of our cur-

rent project called Stairway to elders: bridging space,

time and emotions in their social environment for

wellbeing”, aiming at devising a framework to en-

hance the condition of wellbeing for the aging pop-

ulation, focusing on affective, cognitive, and social

factors.

In this note the project is introduced by outlin-

ing the main research ambition, namely the devel-

opment of computational methods for understand-

ing the affective state and the socialization attitude

of older people. This will be grounded on multi-

modal observations acquired over medium/long tem-

poral frames. In order to reduce the Hawthorne effect,

that is the participant’s behaviour alteration because

of the awareness of participating in an investigation

(Jones, 1992), the experiments will be carried out in

unconstrained ecological environments, targeting the

elderly common activities. In cascade, we will ex-

plore speciﬁc interventions on the environment able

to affect positively on the individual mood.

The paper is organized as follows: in Sec. 2, we

present a brief overview of the most recent Emotion-

aware Ambient intelligence; in Sec. 3, we provide

a detailed outline of our research objectives. Then,

in Sec. 4, we illustrate the expected results, and in

Sec. 5 we draw some conclusions and provide ﬁnal

considerations.

2 RELATED WORKS

Most of the investigations conducted in affective com-

puting have been performed in reductive contexts,

which may limit their applicability and may generate

biased outcomes (Jallais and Gilet, 2010; Zhang et al.,

2014). On the contrary, our project is grounded on the

idea of implementing it in ecological environments,

enriching the so called Ambient Intelligence (AmI)

with actuators, so as to make it responsive to human

needs and emotions. This concept is often referred to

as Emotion-aware Ambient intelligence (AmE) (Zhou

et al., 2007).

Some proofs of concept have already been pre-

sented in the literature, even though they do not pro-

pose speciﬁc computational models for elderly well-

being. In (Fern

andez-Caballero et al., 2016), for ex-

ample, the authors propose a generic, open and adapt-

able AmE architecture for emotion detection and reg-

ulation. This architecture should capture and inte-

grate physiological signals, face expression and body

movement and estimate the emotional state of the

monitored person. A pleasant state of mood should

then be induced by music and colour/light actuation.

In (Mano et al., 2016) a three layer AmE architecture

is introduced relying upon the following: simple and

dedicated devices that act as sensors collecting the in-

formation about the patients health/emotional status;

a decision maker with more powerful computing re-

sources; an actuator, that should be a simple alert to

caregivers. In (Rodrigues and Pereira, 2018), a uni-

ﬁed model promoting the wellbeing of the elderly liv-

ing at home is proposed. It takes into account three

aspects concerning the wellbeing: health, activity, and

emotions. As for emotions, they adopt a smiling count

to determine the happiness of people. Such measure

would then be integrated with personalized health and

activity coefﬁcients so that the AmE could activate

proper actuators (e.g. to propose an activity, turn on a

display, or alert the caregivers).

Overall, the most crucial lack we observe in ex-

isting solutions is that the detection of affective state

and social attitude - namely, the core of AmE - is often

obtained referring to general purpose classiﬁers taken

on-the-shelf from generic affective computing tools

(Grossi et al., 2019). Here and in our project we point

out the importance of learning techniques relying on

realistic data acquired in ecological conditions and in-

volving older people, allowing to tailor a solution for

this peculiar population.

3 RESEARCH FLOW

Goal of this project is to design a framework en-

dowing it with the automatic assessment of affective

state and social attitude in the elderly population; the

model should include a feedback to allow possible

corrective interventions to spatio-temporal environ-

ment, in case low-mood or negative attitude are ob-

served (see Fig. 1).

Speciﬁcally, the project concerns three main ob-

jectives: (1) Identiﬁcation of key contexts and the

consequent acquisition of ecological data through

multimodal and non-invasive sensors (e.g., cameras,

Stairway to Elders: Bridging Space, Time and Emotions in Their Social Environment for Wellbeing

549

and wristbands), covering an appropriate space and

time, and promoting comfort during the acquisition.

(2) Deﬁnition and implementation of models for the

quantitative evaluation of the affective state and the

social attitude of an individual in an ideal continuum

of the emotional space, with particular attention to

subtle changes (as an episode of anger or frustration

due to a temporary failure in a performed activity) or,

conversely, slower variations that can be only appre-

ciated through monitoring over longer time periods

(e.g. an increasing sense of unease due to an un-

pleasant social context). (3) Provision of feedback,

with mild interventions on the space and time dimen-

sions (to improve comfort and stimulate positive feel-

ings), and an eventual customization of the interven-

tion, with respect to a speciﬁc person or a particular

mood-state.

Models of

Emotion &

Social

Understanding

MOOD

Social

Attitude

Automatic Framework

Space/Time

Environment

Figure 1: Workﬂow of the proposed system: the environ-

ment inﬂuences both emotions and social interactions of el-

derly people; social signals are acquired and an automatic

framework provides the social attitude level and the emo-

tional state so as to guide contingent interventions.

In the following we comment on each proposed

objective.

3.1 Key Contexts

The identiﬁcation of suitable monitoring scenarios is

a key element to spot possible frailties. As detailed in

the following, in our project we will take into account

normal activities of daily living in a homely environ-

ment, one-to-one laboratory sessions aimed at stimu-

lating motor or cognitive skills, and social activities

that require either interaction among elders or with

younger people (e.g. family members, operators, vis-

itors). Such a variety of contexts is possible thank to

the collaboration with different stakeholders allowing

us to have available multiform environments, several

live styles (in community or in private ﬂats), and dif-

ferent degrees of frailty. In particular, we will involve

elders attending a university of the third age (minimal

frailty), elders living in a rest home (with multiform

level of frailties), and others living in ﬂats annex to a

hospital (various level of frailties).

3.1.1 Watching Television

For aged people living alone, watching TV often rep-

resents one of the main daily activities, thus deserving

a speciﬁc study to make this time fruitful in giving a

positive feeling to elders, and to investigate the factors

to favor this. For example, we will compare the effect

of conducting such activity alone or in a group and

the effect of tuning the video speed to users percep-

tual and cognitive abilities. Assessing the core affect

of the TV watcher (via physiological data, face ex-

pression, and questionnaires) will allow us to derive

his/her mood, and consequently to intervene on the

fruition modality to induce a sense of fulﬁlment.

3.1.2 Free Time

This context concerns all the unconstrained scenarios,

where elders conduct activities moving in the space

freely (e.g. reading, playing board games, having a

conversation or taking a rest, cooking). This context is

suitable to reveal the way elders spontaneously relate

with the environment around them, their psychologi-

cal stability and, if conducted in a social context, their

socialization attitude. In this task, we will investi-

gate the physiological data and body movement, more

reliable than other cues given the unconstrained sce-

nario. Speciﬁc attention will be posed on designing

motion representations able to embed emotional fea-

tures, bridging well known formulations describing

biological motion and its regularities (Noceti et al.,

2017) with the popular valence-arousal model (Rus-

sell, 2003).

As for the socialization attitude, it will be in-

vestigated through a videobased approach analysing

groups formation and the mutual relation between

people in the group, exploiting pose (Cao et al., 2018)

and gaze direction (Recasens et al., 2015; Dias et al.,

2019) estimation. The derived information will guide

automatic changes in the environmental conditions

(light and music), in order to facilitate emotion reg-

ulation.

3.1.3 Laboratory

This group of experiments will address speciﬁcally

community life, where operators propose and conduct

individual or group activities to stimulate either cog-

nitive/affective, motor or social aspects. Information

ICPRAM 2020 - 9th International Conference on Pattern Recognition Applications and Methods

550

gathered within-lab, concerning the positive or neg-

ative impact on subjects, provides an essential feed-

back to the operators for understanding whether the

stimulation is effective or not, and in which measure.

In such a controlled setting, we will gather physiolog-

ical signals, body movements, and, if possible, face

expressions. Also, in this task, we will investigate the

mutual pose and gaze direction, ﬁnalized to explore

the degree of involvement of older people in the ac-

tivity, and also their mutual collaboration.

3.1.4 Video Call

This task will concern the video call between elders

and relatives or friends. A positive communication

will concern both emotional and cognitive aspects.

On the one hand, monitoring the emotion aroused by

accomplishing this task will allow to assess the psy-

chological stability of elders. On the other hand, we

are interested in exploring the possibility of tuning the

video ﬂow according to the preferences of both speak-

ers and listeners (typically slower for elders), for ex-

ample by introducing a buffer in the communication

or by slowing down pitchcompensated video commu-

nication.

3.2 Emotion and Social Attitude

Quantiﬁcation

This module aims at determining in a continuum the

emotional state and social attitude of elders, by using

multimodal information, i.e., facial expression, phys-

iological signals, and motion quality. This is useful

both when an individual is alone, to allow automatic

environment changes or to give alert to relatives, and

when he/she lives in a community, where the pro-

fessional operators might miss subtle discomfort or

poorly expressed emotions or needs. This objective

will be accomplished facing three tasks, namely the

emotion recognition, the social attitude estimation,

and the analysis concerning video tuning.

3.2.1 Core Affect Learning from Multimodal

Signalling

In order to exploit, when possible, the strength of

multimodal social signals, a probabilistic framework

(Boccignone et al., 2018) will be exploited, suitable

for dealing with the variety of input signals and re-

lated uncertainties. Such a model accounts for multi-

modal signal dynamics in terms of trajectories unfold-

ing in a continuous, central affect state-space (cfr.,

(Anderson and Adolphs, 2014) for an in-depth dis-

cussion) akin to the well known valence-arousal core

affect space proposed by Russell (Russell, 2003) (see

Fig. 2).

Video Analysis

of face/gesture

Physiological

signal analysis

Inference

Mapping

Model of Emotion Understanding

Valence Score

Arousal Score

Figure 2: Workﬂow of the emotion sensing and understand-

ing: social signals are captured and mapped into the va-

lence/arousal space by a Bayesian deep model.

We will apply the model to a combination of phys-

iological and facial cues, and consider the applicabil-

ity to motion features and fullbody expressions.

The model is to be learnt in a supervised setting,

from an appropriate amount of data, labelled (Boc-

cignone et al., 2017) by domain experts according to

the perceived emotion, either producing a personal-

ized model for domestic use, or a general one for com-

munities.

3.2.2 Social Attitude

Social attitude will be derived analysing gaze direc-

tion (Dias et al., 2019; Cuculo et al., 2018), body

pose (Cao et al., 2018), and movements of people

forming a group. The literature in expressive motion

analysis is largely based on motion capture systems,

with fewer results on video analysis. Video signals are

instead easier to adopt (cheaper sensors, lower setup

requirements) and they provide rich sources of het-

erogeneous information. For this reason, they will be

considered as a main source of data in our project;

this is feasible thanks to the availability of recently

proposed pose estimation methods, which allow us

to obtain information on the 2D or 3D body posture

from video. We will ﬁrst evaluate the applicability of

the existing algorithms to the speciﬁc application do-

main, possibly improving, if necessary, the pose es-

timation precision. We will also incorporate a tem-

poral analysis of the pose, based on dynamic ﬁlters.

Then, we will exploit these measurements to obtain

an estimate of the quantity and quality of social inter-

actions among subjects, as a further cue of the over-

all affective state. As already stated, the motion fea-

tures and full-body expressions will also provide a

ﬁrst coarse estimate of the emotional state, (extend-

ing (Piana et al., 2016)) that could be used as an ad-

ditional input for the multimodal model described in

Sec. 3.2.1.

Stairway to Elders: Bridging Space, Time and Emotions in Their Social Environment for Wellbeing

551

3.2.3 Video Tuning

This task aims at deﬁning the optimal tuning to video

speed and rhythms, and at ﬁnding possible relations

with individual sensorimotor rhythms, which in the

elderly may be rather slowed (Salthouse, 1996). We

will investigate video speed preferences at the per-

ceptual and emotional levels with psychophysical

and psychophysiological methods (Rossi et al., 2018;

Mackersie and Kearney, 2017; deSperati and Baud-

Bovy, 2017; Zuliani et al., 2019; de’Sperati, 2020)

and facial emotion recognition. We will keep into

account individual abilities in the cognitive (time esti-

mation through duration reproduction tasks (Grondin,

2010)) behavioural (through the analysis of speed-

accuracy tradeoff, (Heitz, 2014)) and motor (through

kinematic analysis of simple movements such as

grasping (Bruno et al., 2016)) domains, as well as

contextual factors such as time of the day, arousal,

mood, etc. The expected outcome is the comprehen-

sion of how to optimally regulate video speed (in-

creased speed, decreased speed, adjustable speed con-

trol, no change), which in perspective may lead to a

change in TV and video standards to meet not only

the elder population.

3.3 Intervention Deﬁnition

Given the assessment of emotional state and social at-

titude, several interventions could be considered, all

ﬁnalized to enhance the elders life condition and gain

emotional stability (Jallais and Gilet, 2010; Zhang

et al., 2014). The interventions will be either auto-

matic or guided by operators.

The automatic interventions on the domestic space

will include changes in sensory stimulation such as

light modulation (Cuculo et al., 2015) or music stim-

ulation (Anttonen and Surakka, 2007), the former to

adjust the environment in order to match the relax-

ation/arousing state or to favour interaction, and the

latter to modulate or evoke emotions. Furthermore,

the regulation of the reproduction speed of certain TV

programs, and of the video ﬂow during video call will

induce a sense of fulﬁlment.

Finally, in close collaboration with the stakehold-

ers and the caregivers, we will identify interventions

guided by operators, aiming at improving the emo-

tional state and the sense of inclusion.

4 EXPECTED OUTCOMES AND

SOCIAL IMPACT

Our proposal answers to the emerging question on

how to improve the quality of life of elders, (Martini

et al., 2018), considering in particular their emotional

and social spheres. Overall, the research will promote

the design of protocols for the longterm, automatic,

and quantitative assessment of the emotional well-

being level of aging population, to complement the

traditional evaluations performed by the physicians.

Also, it will explore the possibility of producing adap-

tive feedback or interventions to the surrounding en-

vironments, to control and improve such emotional

state. We identify in particular the following general

aspects the project can bring impact on:

• the research will provide a stream of analysis to

evaluate the overall emotional status of an aged

person.

• It will introduce strategies to “represent” the per-

sonal perception of space and time of an indi-

vidual, with consequent understanding of his/her

sense of loneliness and social attitude.

• The framework will be designed to be reliable in

social care facilities as well as private homes, pro-

viding an adaptive solution to generic settings and

a variety of social contexts.

• An affective dataset focused on older people will

be released, including physiological data, facial

expression descriptors (e.g. Action Units (Ekman

and Friesen, 1971)), and emotions annotations on

the valence-arousal 2-dimensional space. This

dataset will be freely available, as benchmark, to

the scientiﬁc community.

5 CONCLUSION

In this paper, we highlighted the emergent issue of

accounting for the psychological wellbeing of elder

people, outlining a proposal to address the hitherto

neglected problem of estimating elderly wellbeing in

ecological or close-to-ecological contexts. Indeed,

besides the physical wellbeing, current health facil-

ities do not actually focus on mental and social as-

pects, that should deserve equal attention.

Under such circumstances, the rationale behind

our proposal is to conceive a well founded frame-

work to automatically assess the affective state in ag-

ing population in order to promote suitable interven-

tion. Namely, the research activity will focus on the

design and validation of methods for continuous de-

tection and analysis of the emotional wellbeing and

ICPRAM 2020 - 9th International Conference on Pattern Recognition Applications and Methods

552

social interaction level of aging people, from multi-

modal sources of information in an ecological set-

ting. Meanwhile, monitoring activity will be paired

with strategic interventions deployed from the envi-

ronment in order to induce positive emotions, in case

of lowmood detection, and to improve the quality of

social interactions.

ACKNOWLEDGEMENTS

This work has been supported by Fondazione Cariplo,

through project Stairway to elders: bridging space,

time and emotions in their social environment for

wellbeing, grant no. 2018-0858.

REFERENCES

Anderson, D. J. and Adolphs, R. (2014). A framework for

studying emotions across species. Cell, 157(1):187–

200.

Anttonen, J. and Surakka, V. (2007). Music, heart rate, and

emotions in the context of stimulating technologies. In

International Conference on Affective Computing and

Intelligent Interaction, pages 290–301. Springer.

Boccignone, G., Conte, D., Cuculo, V., DAmelio, A.,

Grossi, G., and Lanzarotti, R. (2018). Deep construc-

tion of an affective latent space via multimodal enact-

ment. IEEE Transactions on Cognitive and Develop-

mental Systems, 10(4):865–880.

Boccignone, G., Conte, D., Cuculo, V., and Lanzarotti, R.

(2017). Amhuse: a multimodal dataset for humour

sensing. In Proceedings of the 19th ACM Interna-

tional Conference on Multimodal Interaction, pages

438–445. ACM.

Bruno, N., Uccelli, S., Viviani, E., and de’Sperati, C.

(2016). Both vision-for-perception and vision-for-

action follow weber’s law at small object sizes, but

violate it at larger sizes. Neuropsychologia, 91:327–

334.

Cao, Z., Hidalgo, G., Simon, T., Wei, S.-E., and Sheikh,

Y. (2018). Openpose: realtime multi-person 2d pose

estimation using part afﬁnity ﬁelds. arXiv preprint

arXiv:1812.08008.

Cuculo, V., DAmelio, A., Lanzarotti, R., and Boccignone,

G. (2018). Personality gaze patterns unveiled via auto-

matic relevance determination. In Federation of Inter-

national Conferences on Software Technologies: Ap-

plications and Foundations, pages 171–184. Springer.

Cuculo, V., Lanzarotti, R., and Boccignone, G. (2015). The

color of smiling: computational synaesthesia of facial

expressions. In International Conference on Image

Analysis and Processing, pages 203–214. Springer.

de’Sperati, C. (2020). Cortical resonance to visible and in-

visible visual rhythms. Brain Sciences, Submitted.

deSperati, C. and Baud-Bovy, G. (2017). Low perceptual

sensitivity to altered video speed in viewing a soccer

match. Scientiﬁc reports, 7(1):15379.

Dias, P. A., Malafronte, D., Medeiros, H., and Odone, F.

(2019). Gaze estimation for assisted living environ-

ments. In WACV.

Ekman, P. and Friesen, W. V. (1971). Constants across cul-

tures in the face and emotion. Journal of personality

and social psychology, 17(2):124.

Fern

andez-Caballero, A., Mart

ınez-Rodrigo, A., Pastor,

J. M., Castillo, J. C., Lozano-Monasor, E., L

opez,

M. T., Zangr

oniz, R., Latorre, J. M., and Fern

andez-

Sotos, A. (2016). Smart environment architecture for

emotion detection and regulation. J. Biomed. Inform.,

64:55–73.

Fried, L. P., Tangen, C. M., Walston, J., Newman, A. B.,

Hirsch, C., Gottdiener, J., Seeman, T., Tracy, R., Kop,

W. J., Burke, G., et al. (2001). Frailty in older adults:

evidence for a phenotype. The Journals of Gerontol-

ogy Series A: Biological Sciences and Medical Sci-

ences, 56(3):M146–M157.

Giannakouris, K. et al. (2008). Ageing characterises the

demographic perspectives of the european societies.

Statistics in focus, 72:2008.

Grondin, S. (2010). Timing and time perception: a re-

view of recent behavioral and neuroscience ﬁndings

and theoretical directions. Attention, Perception, &

Psychophysics, 72(3):561–582.

Grossi, G., Lanzarotti, R., Napoletano, P., Noceti, N., and

Odone, F. (2019). Positive technology for elderly

well-being: A review. Pattern Recognition Letters.

Hawkins, B. A. (2005). Aging well: toward a way of life

for all people. Preventing chronic disease, 2(3).

Heitz, R. P. (2014). The speed-accuracy tradeoff: history,

physiology, methodology, and behavior. Frontiers in

neuroscience, 8:150.

Jallais, C. and Gilet, A.-L. (2010). Inducing changes

in arousal and valence: Comparison of two mood

induction procedures. Behavior research methods,

42(1):318–325.

Jones, S. R. (1992). Was there a hawthorne effect? Ameri-

can Journal of sociology, 98(3):451–468.

Mackersie, C. L. and Kearney, L. (2017). Autonomic ner-

vous system responses to hearing-related demand and

evaluative threat. American journal of audiology,

26(3S):373–377.

Mano, L. Y., Faic¸al, B. S., Nakamura, L. H., Gomes, P. H.,

Libralon, G. L., Meneguete, R. I., Filho, G. P., Gi-

ancristofaro, G. T., Pessin, G., Krishnamachari, B.,

and Ueyama, J. (2016). Exploiting IoT technologies

for enhancing Health Smart Homes through patient

identiﬁcation and emotion recognition. Comput. Com-

mun., 89:178–190.

Martini, C., Barla, A., Odone, F., Verri, A., Cella, A., Rol-

landi, G. A., and Pilotto, A. (2018). Data-driven con-

tinuous assessment of frailty in older people. Frontiers

in Digital Humanities, 5:6.

Noceti, N., Odone, F., Sciutti, A., and Sandini, G. (2017).

Exploring biological motion regularities of human ac-

Stairway to Elders: Bridging Space, Time and Emotions in Their Social Environment for Wellbeing

553

tions: A new perspective on video analysis. ACM

Transactions on Applied Perception (TAP), 14(3):21.

Piana, S., Staglian

o, A., Odone, F., and Camurri, A. (2016).

Adaptive body gesture representation for automatic

emotion recognition. ACM Transactions on Interac-

tive Intelligent Systems (TiiS), 6(1):6.

Recasens, A., Khosla, A., Vondrick, C., and Torralba, A.

(2015). Where are they looking? In Advances in

Neural Information Processing Systems, pages 199–

207.

Rodrigues, N. and Pereira, A. (2018). A user-centred

well-being home for the elderly. Applied Sciences,

8(6):850.

Rossi, F., Montanaro, E., and deSperati, C. (2018). Speed

biases with real-life video clips. Frontiers in Integra-

tive Neuroscience, 12:11.

Russell, J. A. (2003). Core affect and the psychologi-

cal construction of emotion. Psychological review,

110(1):145.

Salthouse, T. A. (1996). The processing-speed theory of

adult age differences in cognition. Psychological re-

view, 103(3):403.

Scanaill, C. N., Carew, S., Barralon, P., Noury, N., Lyons,

D., and Lyons, G. M. (2006). A review of approaches

to mobility telemonitoring of the elderly in their liv-

ing environment. Annals of biomedical engineering,

34(4):547–563.

Yared, R. and Abdulrazak, B. (2016). Ambient technology

to assist elderly people in indoor risks. Computers,

5(4):22.

Zhang, X., Yu, H. W., and Barrett, L. F. (2014). How does

this make you feel? a comparison of four affect induc-

tion procedures. Frontiers in psychology, 5:689.

Zhou, J., Yu, C., Riekki, J., and K

arkk

ainen, E. (2007). Ame

framework: a model for emotion-aware ambient intel-

ligence. In Proc. of the Int. Conf. on Affective Com-

puting and Intelligent Interaction, page 45.

Zuliani, E., Caputi, M., Scaini, S., and de’Sperati, C.

(2019). Videos look faster as children grow up: sense

of speed and impulsivity throughout primary school.

Journal of experimental child psychology, 179:190–

211.

ICPRAM 2020 - 9th International Conference on Pattern Recognition Applications and Methods

554