Submitted by Valquíria Barbieri (kikibarbi@hotmail.com) on 2017-11-08T19:23:46Z No. of bitstreams: 1 DISS_2012_Nelsindo de Moraes da Silva.pdf: 2248126 bytes, checksum: 62ffc05f3ca16c93b568184f75bfdd61 (MD5) / Approved for entry into archive by Jordan (jordanbiblio@gmail.com) on 2017-12-15T14:49:57Z (GMT) No. of bitstreams: 1 DISS_2012_Nelsindo de Moraes da Silva.pdf: 2248126 bytes, checksum: 62ffc05f3ca16c93b568184f75bfdd61 (MD5) / Made available in DSpace on 2017-12-15T14:49:57Z (GMT). No. of bitstreams: 1 DISS_2012_Nelsindo de Moraes da Silva.pdf: 2248126 bytes, checksum: 62ffc05f3ca16c93b568184f75bfdd61 (MD5) Previous issue date: 2012-05-17 / A presente pesquisa se propõe à descrição do processo de criação musical envolvido em Instâncias, buscando alargar as discussões em torno da relação entre paisagem sonora e composição. A construção da obra teve como referência os objetos sonoros (SCHAEFFER, 1993) presentes na paisagem sonora (SCHAFER, 1991) urbana de Cuiabá, percebida como imagens geradoras (SALLES, 2004) à criação. Para tanto, foram realizadas gravações de áudio em quatro ambientes correlacionados com aspectos do existir humano ─ elaborados a partir da leitura dos pensadores do contemporâneo como Zygmunt Bauman, Michel Maffesoli, Mircea Eliade, Marc Augé e Massimo Canevacci ─, respectivamente: o centro urbano (praças e calçadões), referente à mobilidade; o setor de produção de uma indústria, na dimensão do trabalho; a praça de alimentação de um shopping center, ligada ao prazer; e templos religiosos, conectados à espiritualidade. Assim, considera-se que a textura sonora, emanada destas realidades, manifesta-se como síntese das dimensões que permeiam o existir. Para o tratamento dos dados (processo de escuta, registro e classificação) foi utilizada a teoria de Pierre Schaeffer (1993) acerca da escuta reduzida e do objeto sonoro. Os objetos foram registrados e classificados tendo como sistema de referências a notação musical tradicional; expressados em motes (melódicos e rítmicos), constituindo-se assim como premissas ao processo criativo. Instâncias foi composta para uma orquestra de câmara, dividida em quatro movimentos, são eles: Mobilidade (1º Movimento), Trabalho (2º Movimento), Hedônico (3º Movimento) e Espiritual (4º Movimento). / This research aims to describe the process of musical creation of Instâncias (Instances), trying to extend the matter for soundscape and composition relations. The musical construction has had as references the sound objects (SCHAEFFER, 1993) present in the soundscape (SCHAFER, 1991) from the city of Cuiabá, perceived as generating images (SALLES, 2004) for creating. To this objective, audio recordings have done in four environments correlated with aspects of human existence ─ developed from the reading of contemporary thinkers such as Zygmunt Bauman, Michel Maffesoli, Mircea Eliade, Marc Augé and Massimo Canevacci – respectively: the urban center (squares and sidewalks), concerning to the mobility; the production sector of an industry for the work dimension; the food court of a shopping center, linked it to pleasure and, at last, religious temples, connected them to spirituality. Thus, it is considered that the sound texture, emanating from these realities, manifests itself like a synthesis of the dimensions, which permeate the existence. For data processing (listening process, registration and classification), the theory of Pierre Schaeffer (1993), about the reduced listening and the sound object, has applied. The objects have recorded and classified on the traditional musical notation as references system; expressed in motifs (melodic and rhythmic), establishing them as postulations of creative process. Instâncias was composed for a chamber orchestra, in four movements, they are: Mobility (1st Movement), Work (2nd Movement), Hedonic (3rd Movement) and Spiritual (4th Movement).

The usage of image classification in various industries has grown significantly in recentyears. There are however challenges concerning the data used to train such models. Inmany cases the data used in training is often difficult and expensive to obtain. Furthermore,dealing with image data may come with additional problems such as privacy concerns. Inrecent years, synthetic image generation models such as Stable Diffusion has seen signifi-cant improvement. Solely using a textual description, Stable Diffusion is able to generate awide variety of photorealistic images. In addition to textual descriptions, other condition-ing models such as ControlNet has enabled the possibility of additional grounding infor-mation, such as canny edge and segmentation images. This thesis investigates if syntheticimages generated by Stable Diffusion can be used effectively in training an image classifier.To find the most effective method for generating training data, multiple conditioning meth-ods are investigated and evaluated. The results show that it is possible to generate high-quality training data using several conditioning techniques. The best performing methodwas using canny edge grounded images to augment already existing data. Extending twoclasses with additional synthetic data generated by the best performing method, achievedthe highest average F1-score increase of 0.85 percentage points compared with a baselinesolely trained on real images.