This thesis investigates the use of sampled CAD models for training and calibrating a semantic segmentation model, RandLA-Net, with the ultimate goal of localizing modules for digital twinning (the process of creating digital twins). A significant contribution is the development of the Random Placement of Component Generator (RPCG), a synthetic dataset generator that randomly places CAD models within scenes while preserving contextual information such as typical height above ground. Training and testing on datasets generated by RPCG demonstrated its ability to recognize class modules in various randomly generated scenes. Various hyperparameters related to the loss function and pre-processing steps were explored to improve RandLA-Net’s generalization to different contextual settings. Notably, using a class-weighted α in the focal loss showed promise in correctly classifying infrequent classes and reducing network overconfidence under domain shifts with similar prior probability distributions. The semantic segmentation results were promising for the RPCG test set, achieving a mean True Positive Rate (mTPR) of 98% and a mean Intersection over Union(mIoU) of 93.6%. However, the performance on a sampled version of a CAD model representing an installation named Undercentral was comparatively lower, with a mTPR of 41.1% and a mIoU of 33.4%, indicating the need for further adaptation to varied contextual environments. Proposed improvements include enhancing RPCG with an occupancy grid to better simulate compact scenes and evaluating different subsampling rates in RandLA-Net’s random sampling layers. For confidence calibration, the thesis finds that averaging multiple Monte Carlo (MC) dropout evaluations effectively reduces network overconfidence and improves model reliability. Although this work addresses only a portion of the overall digital twinning process, it highlights the potential of synthetic data generation in enhancing semantic segmentation models and contributes towards the localization of modules in digital twin creation.

This thesis investigates predicting the Intersection over Union (IoU) in detections made by the face detector YOLO5Face, which is done to use the predicted IoU as a new uncertainty measure. The detections are done on the face dataset WIDER FACE, and the prediction of IoU is made by adding a parallel head to the existing YOLO5Face architecture. Experiments show that the methodology for predicting the IoU used in this thesis does not work and the parallel prediction head fails to predict the IoU and instead resorts to predicting common IoU values. The localisation confidence and classification confidences of YOLO5Face are then investigated to find out which confidence measure is least uncertain and most suitable to use when identifying faces. Experiments show that the localisation confidence is consistently more calibrated than the classification confidence. The classification confidence is then calibrated with respect to the localisation confidence which reduces the Expected Calibration Error (ECE) for classification confidence from 0.17 to 0.01.