RT Dissertation/Thesis
T1 Fear Classification using Affective Computing with Physiological Information and Smart-Wearables
A1 Miranda Calero, José Ángel
AB Among the 17 Sustainable Development Goals proposed within the 2030 Agendaand adopted by all of the United Nations member states, the fifth SDG is a callfor action to effectively turn gender equality into a fundamental human right andan essential foundation for a better world. It includes the eradication of all typesof violence against women. Focusing on the technological perspective, the range ofavailable solutions intended to prevent this social problem is very limited. Moreover,most of the solutions are based on a panic button approach, leaving asidethe usage and integration of current state-of-the-art technologies, such as the Internetof Things (IoT), affective computing, cyber-physical systems, and smart-sensors.Thus, the main purpose of this research is to provide new insight into the design anddevelopment of tools to prevent and combat Gender-based Violence risky situationsand, even, aggressions, from a technological perspective, but without leaving asidethe different sociological considerations directly related to the problem. To achievesuch an objective, we rely on the application of affective computing from a realistpoint of view, i.e. targeting the generation of systems and tools capable of being implementedand used nowadays or within an achievable time-frame. This pragmaticvision is channelled through: 1) an exhaustive study of the existing technologicaltools and mechanisms oriented to the fight Gender-based Violence, 2) the proposalof a new smart-wearable system intended to deal with some of the current technologicalencountered limitations, 3) a novel fear-related emotion classification approachto disentangle the relation between emotions and physiology, and 4) the definitionand release of a new multi-modal dataset for emotion recognition in women.Firstly, different fear classification systems using a reduced set of physiological signals are explored and designed. This is done by employing open datasets togetherwith the combination of time, frequency and non-linear domain techniques. Thisdesign process is encompassed by trade-offs between both physiological considerationsand embedded capabilities. The latter is of paramount importance due tothe edge-computing focus of this research. Two results are highlighted in this firsttask, the designed fear classification system that employed the DEAP dataset dataand achieved an AUC of 81.60% and a Gmean of 81.55% on average for a subjectindependentapproach, and only two physiological signals; and the designed fearclassification system that employed the MAHNOB dataset data achieving an AUCof 86.00% and a Gmean of 73.78% on average for a subject-independent approach,only three physiological signals, and a Leave-One-Subject-Out configuration. A detailedcomparison with other emotion recognition systems proposed in the literatureis presented, which proves that the obtained metrics are in line with the state-ofthe-art.Secondly, Bindi is presented. This is an end-to-end autonomous multimodal systemleveraging affective IoT throughout auditory and physiological commercial off-theshelfsmart-sensors, hierarchical multisensorial fusion, and secured server architectureto combat Gender-based Violence by automatically detecting risky situationsbased on a multimodal intelligence engine and then triggering a protection protocol.Specifically, this research is focused onto the hardware and software design of one ofthe two edge-computing devices within Bindi. This is a bracelet integrating threephysiological sensors, actuators, power monitoring integrated chips, and a System-On-Chip with wireless capabilities. Within this context, different embedded designspace explorations are presented: embedded filtering evaluation, online physiologicalsignal quality assessment, feature extraction, and power consumption analysis.The reported results in all these processes are successfully validated and, for someof them, even compared against physiological standard measurement equipment.Amongst the different obtained results regarding the embedded design and implementationwithin the bracelet of Bindi, it should be highlighted that its low powerconsumption provides a battery life to be approximately 40 hours when using a 500mAh battery.Finally, the particularities of our use case and the scarcity of open multimodal datasets dealing with emotional immersive technology, labelling methodology consideringthe gender perspective, balanced stimuli distribution regarding the targetemotions, and recovery processes based on the physiological signals of the volunteersto quantify and isolate the emotional activation between stimuli, led us to the definitionand elaboration of Women and Emotion Multi-modal Affective Computing(WEMAC) dataset. This is a multimodal dataset in which 104 women who neverexperienced Gender-based Violence that performed different emotion-related stimulivisualisations in a laboratory environment. The previous fear binary classificationsystems were improved and applied to this novel multimodal dataset. For instance,the proposed multimodal fear recognition system using this dataset reports up to60.20% and 67.59% for ACC and F1-score, respectively. These values represent acompetitive result in comparison with the state-of-the-art that deal with similarmulti-modal use cases.In general, this PhD thesis has opened a new research line within the research groupunder which it has been developed. Moreover, this work has established a solid basefrom which to expand knowledge and continue research targeting the generation ofboth mechanisms to help vulnerable groups and socially oriented technology.
YR 2022
FD 2022-04
LK https://hdl.handle.net/10016/35848
UL https://hdl.handle.net/10016/35848
LA eng
NO Mención Internacional en el título de doctor
DS e-Archivo
RD 1 jul. 2024