In Depth Bayesian Semantic Scene Completion

TL;DR

This paper introduces a Bayesian CNN for semantic scene completion that not only predicts 3D segmentation but also quantifies uncertainty, outperforming standard CNNs especially in complex and unseen scenarios.

Contribution

The paper presents a novel Bayesian CNN model for semantic scene completion that incorporates uncertainty estimation, improving performance over standard CNNs in complex and unseen data conditions.

Findings

Bayesian CNN performs equal or better on MNIST for unseen digits.

Bayesian approach yields better calibrated scores and uncertainty estimates.

Outperforms standard CNN in semantic scene completion on SUNCG dataset.

Abstract

This work studies Semantic Scene Completion which aims to predict a 3D semantic segmentation of our surroundings, even though some areas are occluded. For this we construct a Bayesian Convolutional Neural Network (BCNN), which is not only able to perform the segmentation, but also predict model uncertainty. This is an important feature not present in standard CNNs. We show on the MNIST dataset that the Bayesian approach performs equal or better to the standard CNN when processing digits unseen in the training phase when looking at accuracy, precision and recall. With the added benefit of having better calibrated scores and the ability to express model uncertainty. We then show results for the Semantic Scene Completion task where a category is introduced at test time on the SUNCG dataset. In this more complex task the Bayesian approach outperforms the standard CNN. Showing better Intersection over Union score and excels in Average Precision and separation scores.