Transfer Learning via Parameter Regularization for Medical Image Segmentation

Sagie, Nimrod; Greenspan, Hayit; Goldberger, Jacob

doi:10.23919/eusipco54536.2021.9616331

Cited by 2 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They found that L 2 -SP was beneficial in all tested use cases. Sagie et al [26] evaluated L 2 -SP in a supervised task adaptation setting for segmentation. L 2 -SP was not beneficial for ImageNet pre-training.…”

Section: Discussionmentioning

confidence: 99%

“…At present, researchers using pre-trained DL models for digital histopathology can choose from various recently proposed algorithms from CV research [22][23][24]. Some initially successful applications of advanced transfer learning techniques in the medical domain can be found with respect to radiological images [25][26][27][28] or lung sound analysis [29]. However, the wider adoption of these techniques in medicine has not occurred to date.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Systematic Comparison of Task Adaptation Techniques for Digital Histopathology

Sauter,

Lodde,

Nensa

et al. 2023

Bioengineering

View full text Add to dashboard Cite

Due to an insufficient amount of image annotation, artificial intelligence in computational histopathology usually relies on fine-tuning pre-trained neural networks. While vanilla fine-tuning has shown to be effective, research on computer vision has recently proposed improved algorithms, promising better accuracy. While initial studies have demonstrated the benefits of these algorithms for medical AI, in particular for radiology, there is no empirical evidence for improved accuracy in histopathology. Therefore, based on the ConvNeXt architecture, our study performs a systematic comparison of nine task adaptation techniques, namely, DELTA, L2-SP, MARS-PGM, Bi-Tuning, BSS, MultiTune, SpotTune, Co-Tuning, and vanilla fine-tuning, on five histopathological classification tasks using eight datasets. The results are based on external testing and statistical validation and reveal a multifaceted picture: some techniques are better suited for histopathology than others, but depending on the classification task, a significant relative improvement in accuracy was observed for five advanced task adaptation techniques over the control method, i.e., vanilla fine-tuning (e.g., Co-Tuning: P(≫) = 0.942, d = 2.623). Furthermore, we studied the classification accuracy for three of the nine methods with respect to the training set size (e.g., Co-Tuning: P(≫) = 0.951, γ = 0.748). Overall, our results show that the performance of advanced task adaptation techniques in histopathology is affected by influencing factors such as the specific classification task or the size of the training dataset.

show abstract

Section: Discussionmentioning

confidence: 99%