Sökning: "semantisk segmentering"

Visar resultat 6 - 10 av 58 uppsatser innehållade orden semantisk segmentering.

1. 6. Real-time Unsupervised Domain Adaptation

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Marc Botet Colomer; [2023]
   Nyckelord :Unsupervised Domain Adaptation; Real-Time applications; Online Learning; Self-Learning; Semantic Segmentation; Reinforcement Learning; Oövervakad domänanpassning; Realtidsapplikationer; Onlineinlärning; Självinlärning; Semantisk Segmentering; Förstärkningsinlärning;

   Sammanfattning : Machine learning systems have been demonstrated to be highly effective in various fields, such as in vision tasks for autonomous driving. However, the deployment of these systems poses a significant challenge in terms of ensuring their reliability and safety in diverse and dynamic environments. LÄS MER

3. 7. Self-supervised pre-training of an attention-based model for 3D medical image segmentation

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Albert Sund Aillet; [2023]
   Nyckelord :Computer vision; Deep learning; 3D Medical image segmentation; Self-supervised learning; Datorseende; Djupinlärning; 3D Medicinsk bildsegmentering; Självövervakad träning;

   Sammanfattning : Accurate segmentation of anatomical structures is crucial for radiation therapy in cancer treatment. Deep learning methods have been demonstrated effective for segmentation of 3D medical images, establishing the current standard. However, they require large amounts of labelled data and suffer from reduced performance on domain shift. LÄS MER

4. 8. Skyline Delineation for Localization in Occluded Environments : Improved Skyline Delineation using Environmental Context from Deep Learning-based Semantic Segmentation

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Kyle William Coble; [2023]
   Nyckelord :Skyline delineation; Skyline detection; Semantic segmentation; Terrain based navigation; Digital elevation models; Uncrewed surface vessel; Planetary exploration robots; Horisont avgränsning; Horisont upptäckt; Semantisk segmentering; Terrängbaserad navigering; Digitala höjdmodeller; Obemannat ytfartyg; Planetariska utforskningsrobotar;

   Sammanfattning : This thesis addresses the problem of improving the delineation of skylines, also referred to as skyline detection, in occluded and challenging environments where existing skyline delineation methods may struggle or fail. Delineated skylines can be used in monocular camera localization methods by comparing delineated skylines to digital elevation model data to estimate a position based on known terrain. LÄS MER

5. 9. Semi-Supervised Domain Adaptation for Semantic Segmentation with Consistency Regularization : A learning framework under scarce dense labels

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Daniel Morales Brotons; [2023]
   Nyckelord :Domain Adaptation; Semi-Supervised Learning; Semi-Supervised Domain Adaptation; Semantic Segmentation; Consistency Regularization; Domain Adaptation; Semi-Supervised Learning; Semi-Supervised Domain Adaptation; Semantisk Segmentering; Konsistensregularisering;

   Sammanfattning : Learning from unlabeled data is a topic of critical significance in machine learning, as the large datasets required to train ever-growing models are costly and impractical to annotate. Semi-Supervised Learning (SSL) methods aim to learn from a few labels and a large unlabeled dataset. LÄS MER

7. 10. Camera ISP optimization for computer vision tasks performed by deep neural networks

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Zhenghong Xiao; [2023]
   Nyckelord :Computer Vision; Image Signal Processing; DNN; Datorseende; bildsignalbehandling; DNN;

   Sammanfattning : This thesis aims to improve the performance of Deep Neural Networkss (DNNs) in Computer Vision tasks by optimizing the Image Signal Processor (ISP) parameters. The research investigates the use of simulated RAW images and the application of the DRL-ISP (Deep Reinforcement Learning for Image Signal Processor) method to enhance the accuracy and robustness of DNNs. LÄS MER