Flores, Huber, juhendajaEmenike, Hilary SomtochukwuTartu Ülikool. Loodus- ja täppisteaduste valdkondTartu Ülikool. Arvutiteaduse instituut2023-09-052023-09-052021https://hdl.handle.net/10062/91984Waste is a growing concern for global sustainability, affecting everything from economic development to human health. Effective waste management is critical for mitigating the harmful effects of waste accumulation on society. In this thesis, we develop MIDAS, an innovative sensing approach for effective household waste management. MIDAS builds on the fact that people need to touch objects when they are throwing them away and that these interactions result in thermal footprints on the objects’ surface. By examining the dissipation of the respective thermal footprints, it is possible to identify the waste material and support effective sorting and recycling practices. We validate our approach through extensive empirical benchmarks, demonstrating that MIDAS can recognize a wide range of materials and generalize variations in the people interacting with objects. Our solution paves towards improved sustainability by offering an innovative solution for classifying materials and optimizing household waste management practices.engopenAccessAttribution-NonCommercial-NoDerivatives 4.0 Internationalthermal imagingwaste managementmobile computingpervasive computingrecycling solutionsIoTmagistritöödinformaatikainfotehnoloogiainformaticsinfotechnologyThesis