Efficient Ensemble Model Generation for Uncertainty Estimation with Bayesian Approximation in Segmentation
This work addresses the computational burden of ensemble-based uncertainty estimation in segmentation, which is an incremental improvement for researchers and practitioners in medical imaging or autonomous systems.
The paper tackles the computational inefficiency of ensemble methods for uncertainty estimation in segmentation by proposing a stochastic layer selection method to generate ensembles efficiently and a pixel-wise uncertainty loss to improve predictive performance, achieving improved uncertainty information and predictive performance on two datasets.
Recent studies have shown that ensemble approaches could not only improve accuracy and but also estimate model uncertainty in deep learning. However, it requires a large number of parameters according to the increase of ensemble models for better prediction and uncertainty estimation. To address this issue, a generic and efficient segmentation framework to construct ensemble segmentation models is devised in this paper. In the proposed method, ensemble models can be efficiently generated by using the stochastic layer selection method. The ensemble models are trained to estimate uncertainty through Bayesian approximation. Moreover, to overcome its limitation from uncertain instances, we devise a new pixel-wise uncertainty loss, which improves the predictive performance. To evaluate our method, comprehensive and comparative experiments have been conducted on two datasets. Experimental results show that the proposed method could provide useful uncertainty information by Bayesian approximation with the efficient ensemble model generation and improve the predictive performance.