PseudoSegRT: efficient pseudo-labelling for intraoperative OCT segmentation

Abstract

Purpose: Robotic ophthalmic microsurgery has significant potential to help improve the success of challenging procedures and overcome the physical limitations of the surgeon. Intraoperative optical coherence tomography (iOCT) has been reported for the visualisation of ophthalmic surgical manoeuvres, where deep learning methods can be used for real-time tissue segmentation and surgical tool tracking. However, many of these methods rely heavily on labelled datasets, where producing annotated segmentation datasets is a time-consuming and tedious task.

Methods: To address this challenge, we propose a robust and efficient semi-supervised method for boundary segmentation in retinal OCT to guide a robotic surgical system. The proposed method uses U-Net as the base model and implements a pseudo-labelling strategy which combines the labelled data with unlabelled OCT scans during training. After training, the model is optimised and accelerated with the use of TensorRT.

Results: Compared with fully supervised learning, the pseudo-labelling method can improve the generalisability of the model and show better performance for unseen data from a different distribution using only 2% of labelled training samples. The accelerated GPU inference takes less than 1 millisecond per frame with FP16 precision.

Conclusion: Our approach demonstrates the potential of using pseudo-labelling strategies in real-time OCT segmentation tasks to guide robotic systems. Furthermore, the accelerated GPU inference of our network is highly promising for segmenting OCT images and guiding the position of a surgical tool (e.g. needle) for sub-retinal injections.

Keywords: Deep learning, Pseudo-labelling; Real-time OCT segmentation; Robotic microsurgery; Semi-supervised learning.

Fig. 1

Sub-retinal injection bleb formation (left). A retinal OCT image (right), the top image shows a binary segmentation prediction (Tissue & Background), and the bottom image shows the layer segmentation prediction, where internal limiting membrane (ILM), and Bruch’s membrane (BM) are shown

Fig. 2

Flowchart of the proposed PseudoSegRT method (left) and diagram of U-Net (right). The proposed strategy first trains a segmentation network in a supervised fashion on the limited dataset. The trained model is used to generate pseudo-labels on limited unlabelled data. The limited labelled and pseudo-labelled data are then used to continue training the segmentation network while updating the pseudo-labels until early stopping is achieved

Fig. 3

Qualitative comparison of the proposed PseudoSegRT with the fully supervised [11] and Pseudo-Lee [15] methods on unseen healthy and unhealthy scans. Ground-truth tissue boundaries are labelled in green and predicted ones are labelled in yellow