Browsing by Author "Mensah, Dann"

Now showing 1 - 5 of 5

Ghost in the border wall: Exploring the possibility of a post-dystopian narrative in a video installation based on the border policies of the European Union
(2024) Lämsä, Henry
School of Arts, Design and Architecture | Master's thesis
Many speculative portrayals of the near future often present the world in a dystopian or post-apocalyptic light. This dark and pessimistic view of the not-so-distant future is a common theme seen in various forms of storytelling, such as films, books, games, and visual art. The prevalence of this apocalyptic mindset is often rooted in the current state of the world, including ongoing conflicts, atrocities, and environmental concerns, which collectively shape a bleak outlook on what lies ahead. This thesis delves into the possibilities of utilizing video installation as a medium to explore post-dystopian narratives within the realm of speculative fiction. Its objective is to construct a narrative that embodies hope following a dystopian occurrence. The narrative is inspired by the ideological underpinnings of current European Union policies, particularly in relation to the perceived external threat posed by refugees and immigrants seeking entry through the increasingly fortified borders of the Union. The culmination of this study is a video installation titled Ghost in the Border Wall. The video installation utilizes modern 3D software in combination with the generative artificial intelligence tool Stable Diffusion. This tool creates images based on a text prompt. In this research, it is employed to visualize the world from the perspective of a border control AI. The visualization produced by Stable Diffusion is superimposed on traditional 3D animation to craft a visually striking and digitally hallucinatory reality. The thesis includes a key finding that using AI tools to generate material as part of the narrative and world-building process can aid in envisioning potential futures. It's important to note that this doesn't necessarily result in post-dystopian outcomes. Shifting ingrained thought patterns from a dystopian and apocalyptic outlook to a more positive one requires significant effort. Creating a post-dystopian vision is an ongoing process that extends beyond the scope of this thesis and will continue to evolve in the future. The findings also point to more significant and consequential issues related to the biases present in the datasets used to train Stable Diffusion. As a result, ensuring that AI-generated content avoids leaning towards sexist, racist, and misogynistic themes becomes challenging. The representation of objectified female bodies and the depiction of minorities are notably impacted by biased data, making it harder to present non-stereotypical characters. These observations highlight the potential benefits of training AI models using project-specific datasets for this type of artwork.
A Hybrid Generator Architecture for Controllable Face Synthesis
(2023-08-21) Mensah, Dann
Perustieteiden korkeakoulu | Master's thesis
A Hybrid Generator Architecture for Controllable Face Synthesis
(2023-07-23) Mensah, Dann; Kim, Nam Hee; Aittala, Miika; Laine, Samuli; Lehtinen, Jaakko
A4 Artikkeli konferenssijulkaisussa
Modern data-driven image generation models often surpass traditional graphics techniques in quality. However, while traditional modeling and animation tools allow precise control over the image generation process in terms of interpretable quantities - e.g., shapes and reflectances - endowing learned models with such controls is generally difficult. In the context of human faces, we seek a data-driven generator architecture that simultaneously retains the photorealistic quality of modern generative adversarial networks (GAN) and allows explicit, disentangled controls over head shapes, expressions, identity, background, and illumination. While our high-level goal is shared by a large body of previous work, we approach the problem with a different philosophy: We treat the problem as an unconditional synthesis task, and engineer interpretable inductive biases into the model that make it easy for the desired behavior to emerge. Concretely, our generator is a combination of learned neural networks and fixed-function blocks, such as a 3D morphable head model and texture-mapping rasterizer, and we leave it up to the training process to figure out how they should be used together. This greatly simplifies the training problem by removing the need for labeled training data; we learn the distributions of the independent variables that drive the model instead of requiring that their values are known for each training image. Furthermore, we need no contrastive or imitation learning for correct behavior. We show that our design successfully encourages the generative model to make use of the internal, interpretable representations in a semantically meaningful manner. This allows sampling of different aspects of the image independently, as well as precise control of the results by manipulating the internal state of the interpretable blocks within the generator. This enables, for instance, facial animation using traditional animation tools.
Modern Methods for Instance-Level Image Segmentation
(2018-04-25) Mensah, Dann
Perustieteiden korkeakoulu | Bachelor's thesis
Sonic walks: A listening and walking methodology to reconfigure spatial narratives of place and belonging
(2024) Bogani Amadori, Francesca
School of Arts, Design and Architecture | Master's thesis
People can feel like they belong and do not belong in a particular place. Various factors and individual experiences can influence the feeling of belonging to a particular place. People experience places differently based on their position within specific social interactions, class structures, political ideals, and moral values. Place and belonging, as dynamic processes, are continually (re)shaped by rearrangements of tension and negotiation between diverse local and global power relations, adding complexity to the experiences. They are not only shaped by sensory inputs, such as hearing familiar traffic sounds in the background or smelling the ocean, but also by the relational and emotional dynamics with other bodies and other places. The relational and embodied experiences of place and thus belonging, in turn, are particular to specific environments and a person’s situatedness –a person’s position within specific social interactions, class structures, political ideals, and moral values in a specific situation and place. These embodied and situated experiences allow the body to be a unique source of knowledge. The situated body generates knowledge from encounters and situations as it interacts with intangible forces, such as social norms, and tangible forces, such as being in a crowd or extreme weather environments. These encounters produce forms of knowledge that interconnect the collective aspects of political, social, cultural, and environmental experiences with the personal. These forms of knowledge are inherently subjective and contribute to a nuanced and unique understanding of the dynamics of belonging. Drawing from personal experiences as a foreigner in an unfamiliar place, in this thesis, I explore the dynamics of belonging and not belonging, as well as various experiences of place. The thesis is structured around three artistic experiments that examine how political, cultural, social, and environmental factors shape divergent perceptions of place. Rooted in walk and sound studies, the methodological framework uses concrete approaches such as sound and listening walks facilitated by technology to engage with the environment. Through these artistic experiments, the methods of research progressively developed, transitioning from broad observations to nuanced reflections on specific elements and areas that evoke feelings of belonging. I found that sound, listening, and walking-based research methods offer a unique approach to examining social and cultural environments. By using sound and walk-based research methods, researchers from diverse fields can have a better understanding of issues of belonging and the cultural and social aspects that are present in everyday life but are not immediately evident. Through my thesis, I hope to encourage others to interact with their surroundings in different ways and to explore their relationship with(in) places and the sociocultural structures that shape them.