Transfer learning with multiple convolutional neural networks for soft tissue sarcoma MRI classification

Hermessi, Haithem; Mourali, Olfa; Zagrouba, Ezzeddine

doi:10.1117/12.2522765

Cited by 2 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, medical image domain knowledge has been transferred to a CNN pretrained on natural images by FT the CNN with medical images. 8,9 Pretrained CNNs have been fine-tuned toward the features of one specific image, resulting in better segmentations of that image. 10,11 A patient-specific fine-tuned CNN has learned the specific features of abnormalities and the surrounding healthy tissue of a patient to improve the detection or segmentation in a follow-up scan.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Patient-specific fine-tuning of convolutional neural networks for follow-up lesion quantification

et al. 2020

View full text Add to dashboard Cite

. Purpose: Convolutional neural network (CNN) methods have been proposed to quantify lesions in medical imaging. Commonly, more than one imaging examination is available for a patient, but the serial information in these images often remains unused. CNN-based methods have the potential to extract valuable information from previously acquired imaging to better quantify lesions on current imaging of the same patient. Approach: A pretrained CNN can be updated with a patient’s previously acquired imaging: patient-specific fine-tuning (FT). In this work, we studied the improvement in performance of lesion quantification methods on magnetic resonance images after FT compared to a pretrained base CNN. We applied the method to two different approaches: the detection of liver metastases and the segmentation of brain white matter hyperintensities (WMH). Results: The patient-specific fine-tuned CNN has a better performance than the base CNN. For the liver metastases, the median true positive rate increases from 0.67 to 0.85. For the WMH segmentation, the mean Dice similarity coefficient increases from 0.82 to 0.87. Conclusions: We showed that patient-specific FT has the potential to improve the lesion quantification performance of general CNNs by exploiting a patient’s previously acquired imaging.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Patient-specific fine-tuning of convolutional neural networks for follow-up lesion quantification

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Fine-tuning a pre-trained CNN has the benefit of obtaining good results using only a small data set during the fine-tuning step and has successfully been applied in previous studies. For example, medical image domain knowledge has been transferred to a CNN pre-trained on natural images by finetuning the CNN with medical images 8,9 . Pre-trained CNNs have been fine-tuned towards the features of one specific image, resulting in better segmentations of that image 10,11 .…”

Section: Introductionmentioning

confidence: 99%

Patient-specific fine-tuning of CNNs for follow-up lesion quantification

Jansen¹,

Kuijf²,

Dhara³

et al. 2019

Preprint

View full text Add to dashboard Cite

Convolutional neural network (CNN) methods have been proposed to quantify lesions in medical imaging. Commonly more than one imaging examination is available for a patient, but the serial information in these images often remains unused. CNNbased methods have the potential to extract valuable information from previously acquired imaging to better quantify current imaging of the same patient. A pre-trained CNN can be updated with a patient's previously acquired imaging: patient-specific fine-tuning. In this work, we studied the improvement in performance of lesion quantification methods on MR images after fine-tuning compared to a base CNN. We applied the method to two different approaches: the detection of liver metastases and the segmentation of brain white matter hyperintensities (WMH). The patient-specific fine-tuned CNN has a better performance than the base CNN. For the liver metastases, the median true positive rate increases from 0.67 to 0.85. For the WMH segmentation, the mean Dice similarity coefficient increases from 0.82 to 0.87. In this study we showed that patient-specific fine-tuning has potential to improve the lesion quantification performance of general CNNs by exploiting the patient's previously acquired imaging.

show abstract

Transfer learning with multiple convolutional neural networks for soft tissue sarcoma MRI classification

Cited by 2 publications

References 18 publications

Patient-specific fine-tuning of convolutional neural networks for follow-up lesion quantification

Patient-specific fine-tuning of convolutional neural networks for follow-up lesion quantification

Patient-specific fine-tuning of CNNs for follow-up lesion quantification

Contact Info

Product

Resources

About