Self-Supervised-RCNN for Medical Image Segmentation with Limited Data Annotation

Felfeliyan, Banafshe; Hareendranathan, Abhilash Rakkunedeth; Kuntze, Gregor; Cornell, David H.; Forkert, Nils D.; Jaremko, Jacob L.; Ronsky, Janet L.

doi:10.48550/arxiv.2207.11191

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SSL has been used for medical imaging datasets including US [14] and histology images [15]. Felfeliyan et al [16] applied different distortions to arbitrary areas of unlabeled data and used the improved Mask R-CNN models to predict distortion type and loss information. He et al [17] proposed Masked Autoencoders (SSL-MAE) for ViT backbone models self-supervised pretraining.…”

Section: Self-supervised Learning and Masked Autoencodermentioning

confidence: 99%

“…There is a normalization layer between encoder and decoder in the original SSL-MAE, but its impact on the downstream segmentation task was not notable, therefore it was retained for SSL-MAE pretraining. Influenced by Felfeliyan's work [16], we investigated changing the original mean squared error (MSE) loss function to the root mean squared error (RMSE) + mean absolute error (MAE) loss over the masked patches reconstruction. Other architectures were maintained the same as the SSL-MAE paper.…”

Section: B Ssl-mae Self-supervised Pretrainingmentioning

confidence: 99%

See 1 more Smart Citation

Wrist Ultrasound Segmentation by Deep Learning

Zhou

Rakkunedeth

Keen

et al. 2022

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

Supervised deep learning offers great promise to automate analysis of medical images from segmentation to diagnosis. However, their performance highly relies on the quality and quantity of the data annotation. Meanwhile, curating large annotated datasets for medical images requires a high level of expertise, which is time-consuming and expensive. Recently, to quench the thirst for large data sets with high-quality annotation, self-supervised learning (SSL) methods using unlabeled domain-specific data, have attracted attention. Therefore, designing an SSL method that relies on minimal quantities of labeled data has far-reaching significance in medical images. This paper investigates the feasibility of deploying the Masked Autoencoder for SSL (SSL-MAE) of TransUNet, for segmenting bony regions from children's wrist ultrasound scans. We found that changing the embedding and loss function in SSL-MAE can produce better downstream results compared to the original SSL-MAE. In addition, we determined that only pretraining TransUNet embedding and encoder with SSL-MAE does not work as well as TransUNet without SSL-MAE pretraining on downstream segmentation tasks.

show abstract