Deep co-training for semi-supervised image segmentation

Abstract

In this paper, we aim to improve the performance of semantic image segmentation in a semi-supervised setting where training is performed with a reduced set of annotated images and additional non-annotated images. We present a method based on an ensemble of deep segmentation models. Models are trained on subsets of the annotated data and use non-annotated images to exchange information with each other, similar to co-training. Diversity across models is enforced with the use of adversarial samples. We demonstrate the potential of our method on two challenging image segmentation problems, and illustrate its ability to share information between simultaneously trained models, while preserving their diversity. Results indicate clear advantages in terms of performance compared to recently proposed semi-supervised methods for segmentation.

Semi-Supervised Segmentation of Concrete Aggregate Using Consensus Regularisation and Prior Guidance

“…In this context, Ouali et al (2020) train multiple auxiliary decoders on unlabelled data by enforcing consistency between the class predictions of the main and the auxiliary decoders. Similarly, in (Peng et al, 2020) two segmentation networks are trained via supervision on two disjunct datasets and additionally, by applying a co-learning scheme in which consistent predictions of both networks on unlabelled data are enforced. Another approach based on consensus training is presented by Li et al (2018) and Zhang et al (2020), who use unlabelled data in order to train a segmentation network by encouraging consistent predictions for the same input under different geometric transformations.…”

“…In order to ensure diversity between both decoders, a perturbed version z of the latent representation z with z = F (z), using a perturbation function F (•), is fed to the auxiliary decoder while the uncorrupted representation z is used as input for the main decoder. This procedure of consensus regularisation for semi-supervised segmentation is founded on the rationale that the shared encoder's representation can be enhanced by using the additional training signal obtained from the unlabelled data, acting as additional regularisation on the encoder (Ouali et al, 2020;Peng et al, 2020). Based on the consensus principle (Chao and Sun, 2016), enforcing an agreement between the predictions of multiple decoder branches restricts the parameter search space to cross-consistent solutions and thus, improves the generalisation of the different models.…”

“…In this section, we present a theoretically founded derivation of the limitations behind semi-supervised training using the consensus principle. In an unsupervised training setup based on the consensus principle as described above and as applied in the literature (Ouali et al, 2020;Peng et al, 2020), the training signal is computed based on the discrepancy between the predictions of two or more distinct models. Consequently, knowledge about the reference labels is not required in order to compute the consensus training loss Lcons, which is the reason why also unlabelled data can be leveraged for training.…”

Medical Image Segmentation With Limited Supervision: A Review of Deep Network Models

“…Moreover, it is essential to ensure the different base learners giving different and complementary information about each instance [173], namely, view difference constraint or diversity criterion. Peng et al [174] applied the idea of co-training to semi-supervised segmentation of medical images. Concretely, they trained multiple models on different subsets of the labeled training data and used a common set of unlabeled training images to exchange information with each other.…”

Progressive graph‐based subspace transductive learning for semi‐supervised classification

“…During the last decades, a large family semi‐supervised learning methods have been proposed [8]. For instance, co‐training is an important semi‐supervised learning method, which assumes that data contains multiple conditionally independent feature subsets and data distribution is compatible with the target function of different feature subsets [7, 10, 11]. Kamal et al [12] performed the extensive empirical experiment to compare the co‐training with generative mixture models and expectation maximization (EM) demonstrated that co‐training performed well if the conditional independence assumption indeed holds.…”