Semi-supervised Medical Image Classification with Global Latent Mixing

Gyawali, Prashnna Kumar; Ghimire, Sandesh; Bajracharya, Pradeep; Li, Zhiyuan; Wang, Linwei

doi:10.1007/978-3-030-59710-8_59

Cited by 36 publications

(27 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As such, in our case, any unlabeled data point x u is a random sample from the combination of both datasets X l and X u . This way of guessing the label encourages the model to be consistent across different augmentations [Gyawali et al 2020]. Although following standard practice, we do not propagate gradients by computing the guessed labels; it should be noted that the guessed labels may change, as the segmentation network f is updated over training.…”

Section: Guessing Labelsmentioning

confidence: 99%

Semi-Supervised Learning for Eye Image Segmentation

Chaudhary

Gyawali

Wang

et al. 2021

ACM Symposium on Eye Tracking Research and Applications

Self Cite

View full text Add to dashboard Cite

Recent advances in appearance-based models have shown improved eye tracking performance in difficult scenarios like occlusion due to eyelashes, eyelids or camera placement, and environmental reflections on the cornea and glasses. The key reason for the improvement is the accurate and robust identification of eye parts (pupil, iris, and sclera regions). The improved accuracy often comes at the cost of labeling an enormous dataset, which is complex and time-consuming. This work presents two semi-supervised learning frameworks to identify eye-parts by taking advantage of unlabeled images where labeled datasets are scarce. With these frameworks, leveraging the domain-specific augmentation and novel spatially varying transformations for image segmentation, we show improved performance on various test cases with limited labeled samples. For instance, for a model trained on just 4 and 48 labeled images, these frameworks improved by at least 4.7% and 0.4% respectively, in segmentation performance over the baseline model, which is trained only with the labeled dataset.

show abstract

Section: Guessing Labelsmentioning

confidence: 99%

Semi-Supervised Learning for Eye Image Segmentation

Chaudhary

Gyawali

Wang

et al. 2021

ACM Symposium on Eye Tracking Research and Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…As such, in our case, any unlabeled data point 𝑥 𝑢 is a random sample from the combination of both datasets X 𝑙 and X 𝑢 . This way of guessing the label encourages the model to be consistent across different augmentations [Gyawali et al 2020]. Although following standard practice, we do not propagate gradients by computing the guessed labels; it should be noted that the guessed labels may change, as the segmentation network 𝑓 is updated over training.…”

Section: Ssl 𝐷 : Ssl With Domain-specific Augmentationmentioning

confidence: 99%

Semi-Supervised Learning for Eye Image Segmentation

Chaudhary¹,

Gyawali²,

Wang³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Recent advances in appearance-based models have shown improved eye tracking performance in difficult scenarios like occlusion due to eyelashes, eyelids or camera placement, and environmental reflections on the cornea and glasses. The key reason for the improvement is the accurate and robust identification of eye parts (pupil, iris, and sclera regions). The improved accuracy often comes at the cost of labeling an enormous dataset, which is complex and time-consuming. This work presents two semi-supervised learning frameworks to identify eye-parts by taking advantage of unlabeled images where labeled datasets are scarce. With these frameworks, leveraging the domain-specific augmentation and novel spatially varying transformations for image segmentation, we show improved performance on various test cases. For instance, for a model trained on just 48 labeled images, these frameworks achieved an improvement of 0.38% and 0.65% in segmentation performance over the baseline model, which is trained only with the labeled dataset.

show abstract

“…To address this problem, a naive solution is to simply integrate the off-the-rack semi-supervised learning (SSL) methods onto the federated learning paradigm. However, previous SSL methods are typically designed for centralized training setting [2,10,17,29], which rely heavily on the assumption that the labeled data is accessible to provide necessary assistance for the learning from unlabeled data [1,4]. In consistency-based methods [5,31], for instance, the regularization of perturbation-invariant model predictions needs the synchronous labeled data supervision, in order to obtain the necessary task knowledge to produce reliable model predictions for unlabeled data where the consistency regularization is imposed on.…”

Section: Introductionmentioning

confidence: 99%

Federated Semi-supervised Medical Image Classification via Inter-client Relation Matching

Liu

Yang

Dou

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Federated learning (FL) has emerged with increasing popularity to collaborate distributed medical institutions for training deep networks. However, despite existing FL algorithms only allow the supervised training setting, most hospitals in realistic usually cannot afford the intricate data labeling due to absence of budget or expertise. This paper studies a practical yet challenging FL problem, named Federated Semi-supervised Learning (FSSL), which aims to learn a federated model by jointly utilizing the data from both labeled and unlabeled clients (i.e., hospitals). We present a novel approach for this problem, which improves over traditional consistency regularization mechanism with a new inter-client relation matching scheme. The proposed learning scheme explicitly connects the learning across labeled and unlabeled clients by aligning their extracted disease relationships, thereby mitigating the deficiency of task knowledge at unlabeled clients and promoting discriminative information from unlabeled samples. We validate our method on two large-scale medical image classification datasets. The effectiveness of our method has been demonstrated with the clear improvements over state-of-the-arts as well as the thorough ablation analysis on both tasks 4 .

show abstract

Semi-supervised Medical Image Classification with Global Latent Mixing

Cited by 36 publications

References 7 publications

Semi-Supervised Learning for Eye Image Segmentation

Semi-Supervised Learning for Eye Image Segmentation

Semi-Supervised Learning for Eye Image Segmentation

Federated Semi-supervised Medical Image Classification via Inter-client Relation Matching

Contact Info

Product

Resources

About