Periodic Reporting for period 2 - MindSpaces (MindSpaces - Art-driven adaptive outdoors and indoors design)
Période du rapport: 2020-07-01 au 2022-06-30
People experience the built environment differently according to their social, cultural and economic background. The variety of this experience requires consideration as we try to develop the feelings of attachment and sense of belonging of all users of a particular space or place. Towards this direction, with the use of Virtual Reality, biosensors, Αrtificial Ιntelligence algorithms and simulations, design decisions were taken in relation to a detailed understanding of the users’ preferences and collective behaviour. Based on the above desideratum, MindSpaces seeks to transform the practice of architecture by integrating the full diversity of how people, with different social, cultural and economic backgrounds, experience and behave in the spaces we design by using innovative, artistic - driven, digital technologies. The main objective of MindSpaces was to develop novel art-driven design processes and technologies, which built upon artificial intelligence, multimodal data analysis and fusion algorithms and were augmented by data insights gathered through the collective social behaviour and responses of occupants experiencing dynamic and adaptive environments.
Therefore, MindSpaces is an endeavour that brought together artists and scientists in order to tackle societal and urban city design challenges by leveraging the inspiration of the first and technological innovations of the latter. It addressed three urban and architectural challenges aiming at an overall more humane quality of life, by providing innovative solutions for the creation of (1) attractive, functional outdoors urban spaces that attract visitors, supporting the sense of belonging, social interaction and local commerce, (2) indoors work environments, where people feel comfortable and inspired, increasing their productivity, (3) indoor ambient design for senior people in order to increase reminiscence and decrease loneliness.
Therefore, MindSpaces is an endeavour that brought together artists and scientists in order to tackle societal and urban city design challenges by leveraging the inspiration of the first and technological innovations of the latter. It addressed three urban and architectural challenges aiming at an overall more humane quality of life, by providing innovative solutions for the creation of (1) attractive, functional outdoors urban spaces that attract visitors, supporting the sense of belonging, social interaction and local commerce, (2) indoors work environments, where people feel comfortable and inspired, increasing their productivity, (3) indoor ambient design for senior people in order to increase reminiscence and decrease loneliness.
During the project (M1-M42) MindSpaces achieved all the objectives and reached all milestones set for this reporting period. The user requirements in interior design and in urban scale were defined (MS1), based on which the technical specifications have been set and the toolset architecture has been designed. This architecture was realised by an operational prototype (MS2), which integrated the skeleton versions of the research modules, and served as the basis towards the first prototype (MS3). In MS4 and MS5, we have worked on creating technologies, which targeted PUC in more detail, maintaining a backbone that could support all the services and components. A set of tools and services has been developed for this scope. In MS4, the 2nd version of the MindSpaces toolkit integrated the basic techniques for: a) Instant cities (latent space explorer), b) Computational Design Tool (PUC1), c) ABPS Tools (PUC2), d) Visual Sentiment analysis Tool (PUC2), e) VR tool (PUC2), f) Imageability mobile app (MentalSpaces app), g) Multimodal fusion of sentiment analysis (PUC2), h) Behavioural analysis service (Dialog - Refik Anadol, Maurice Benayoun), i) Emotional and Behavioral Analysis Service. Moreover, an updated standardisation and dissemination report was also performed, an updated data management and self-assessment report and the and the evaluation of the 2nd prototype were also successfully performed. In MS5, the final modules and tools were delivered with the final system and its evaluation, the final activity report and self-assessment, the final dissemination report, the final exploitation report and roadmap to market and the ethics report.
Finally, 37 scientific publications and 2 media publications were published and 9 workshops have been organised, with an average of 25 participants each. The website and the social media managed to increase their audience and usage statistics, which indicated that our online dissemination strategy has proven successful. The website had 25196 visits by 10,06 users with a total amount of 13907 Sessions (2019 - 2022). 15 news items have been published in the final year of the project, 3 Ars Electronica Gardens and 1 big Symposium have been successfully organised, and 2 exhibitions have been curated. MindSpaces continued the close collaboration with artists and user groups (12 users by request, 17 users contacted from partners, 68 users under contact from various domains, as their feedback proved valuable for development of the toolkit and the successful future exploitation of project results.
Finally, 37 scientific publications and 2 media publications were published and 9 workshops have been organised, with an average of 25 participants each. The website and the social media managed to increase their audience and usage statistics, which indicated that our online dissemination strategy has proven successful. The website had 25196 visits by 10,06 users with a total amount of 13907 Sessions (2019 - 2022). 15 news items have been published in the final year of the project, 3 Ars Electronica Gardens and 1 big Symposium have been successfully organised, and 2 exhibitions have been curated. MindSpaces continued the close collaboration with artists and user groups (12 users by request, 17 users contacted from partners, 68 users under contact from various domains, as their feedback proved valuable for development of the toolkit and the successful future exploitation of project results.
MindSpaces provides new tools for the creative industries through the development of 3D and VR design solutions and achieved the following impacts:
• Demonstrates a novel solution for innovative urban and indoors design by directly incorporating artists’ inputs and interventions in the creative process.
• Addresses real-world problems in urban and home settings through the synergy of artists and creatives with SoA technological solutions.
• Proposes a close collaboration scheme and working model between technology and art in order to further unleash the technological potential.
The expected final result in an integrated and flexible toolset revolves around two main axes: (i) improving design in expanding urban environments (PUC1), in modern workplaces (PUC2) and in interior ambient environments for seniors (PUC3); (ii) creating adaptable immersive VR experiences for all the above spaces. MindSpaces pertained to the listed scientific fields and facilitated the implementation of the planned impact.
A summary of the innovations derived by the project are presented below:
Interoperability and semantics: for the enhancement of the MindSpaces system explainability due to the adoption of the semantics in the ontology.
Web crawling and retrieval: a service to monitor tweet posts, timelines, users and objects.
Language analysis and generation: extended lexical and grammatical resources for semantic parsing/generation of English, Spanish, Catalan, French and Greek in the domains of MindSpaces.
Visual Sentiment analysis: creation of a novel dataset for visual sentiment analysis of urban and indoor spaces through the annotation of real images and screenshots from the MindSpaces 3D models in terms of valence and arousal.
Emotional and cognitive analysis: physiological data collection using an immersive VR environment as emotional stimulus.
Multimodal fusion of sentiment analysis: regarding the development of a novel method for multimodal imageability score calculation based on segmentation analysis from visual input, textual analysis and users’ annotations.
Mobile mapping platform and 3D model reconstruction: a platform developed, following the need of a fast and streamlined creation of 3D models for architecture, engineering and construction processes.
VR tool: a Collaborative VR tool, which allows multiple users to be in the same VR experience, interact with each other and share their emotions using emoticons.
MindSpaces Design tool: a new, novel technical framework that supports collaboration in building complex computational design processes.
ABPS tool: a set of several tools developed and integrated into a new data driven design framework for workplaces.
Latent Space Explorer: a tool that works with generative artificial neural networks trained on image data.
• Demonstrates a novel solution for innovative urban and indoors design by directly incorporating artists’ inputs and interventions in the creative process.
• Addresses real-world problems in urban and home settings through the synergy of artists and creatives with SoA technological solutions.
• Proposes a close collaboration scheme and working model between technology and art in order to further unleash the technological potential.
The expected final result in an integrated and flexible toolset revolves around two main axes: (i) improving design in expanding urban environments (PUC1), in modern workplaces (PUC2) and in interior ambient environments for seniors (PUC3); (ii) creating adaptable immersive VR experiences for all the above spaces. MindSpaces pertained to the listed scientific fields and facilitated the implementation of the planned impact.
A summary of the innovations derived by the project are presented below:
Interoperability and semantics: for the enhancement of the MindSpaces system explainability due to the adoption of the semantics in the ontology.
Web crawling and retrieval: a service to monitor tweet posts, timelines, users and objects.
Language analysis and generation: extended lexical and grammatical resources for semantic parsing/generation of English, Spanish, Catalan, French and Greek in the domains of MindSpaces.
Visual Sentiment analysis: creation of a novel dataset for visual sentiment analysis of urban and indoor spaces through the annotation of real images and screenshots from the MindSpaces 3D models in terms of valence and arousal.
Emotional and cognitive analysis: physiological data collection using an immersive VR environment as emotional stimulus.
Multimodal fusion of sentiment analysis: regarding the development of a novel method for multimodal imageability score calculation based on segmentation analysis from visual input, textual analysis and users’ annotations.
Mobile mapping platform and 3D model reconstruction: a platform developed, following the need of a fast and streamlined creation of 3D models for architecture, engineering and construction processes.
VR tool: a Collaborative VR tool, which allows multiple users to be in the same VR experience, interact with each other and share their emotions using emoticons.
MindSpaces Design tool: a new, novel technical framework that supports collaboration in building complex computational design processes.
ABPS tool: a set of several tools developed and integrated into a new data driven design framework for workplaces.
Latent Space Explorer: a tool that works with generative artificial neural networks trained on image data.