Improved-Mask R-CNN: Towards an accurate generic MSK MRI instance segmentation platform (data from the Osteoarthritis Initiative)

Felfeliyan, Banafshe; Hareendranathan, Abhilash Rakkunedeth; Kuntze, Gregor; Jaremko, Jacob L.; Ronsky, Janet L.

doi:10.1016/j.compmedimag.2022.102056

Cited by 17 publications

(13 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MaskRCNN [1] using probabilistic partial labels. 2) Creation of probabilistic labels based on multi-rater scoring, MRI spatial redundancy, and tissue/lesion characteristics, as well as considering image contextual information.…”

Section: ) Exploration Of Training Of An Improved Version Ofmentioning

confidence: 99%

“…We used the IMaskRCNN [1] as the baseline deep learning model for training and evaluation of the proposed extensions. The IMaskRCNN model is an improved version of the wellknown instance segmentation model Mask RCNN [31] that improves the segmentation accuracy around object boundaries by adding a skip connection and an extra encoder layer to the mask segmentation head (inspired by U-net architecture) [31].…”

Section: B Modelmentioning

confidence: 99%

“…1. Generating soft pseudo-GT for BML pipeline segmentation using the IMaskRCNN trained on registered data from our previous work [1]. In the following, we explain how we prepared BML annotations for training.…”

Section: A Data and Labelsmentioning

confidence: 99%

“…Automatic or semi-automatic clinical features and biomarker measurements based on deep learning (DL) are useful for longitudinal assessment of medical images. DL-based techniques using Convolutional Neural Networks (CNN) have shown great success in tissue and pathology detection and segmentation [1], [2]. The accuracy of deep learning methods is highly dependent on the quality of the training data and corresponding ground truth labels.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Weakly Supervised Medical Image Segmentation With Soft Labels and Noise Robust Loss

Felfeliyan¹,

Hareendranathan²,

Kuntze³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Recent advances in deep learning algorithms have led to significant benefits for solving many medical image analysis problems. Training deep learning models commonly requires large datasets with expert-labeled annotations. However, acquiring expert-labeled annotation is not only expensive but also is subjective, error-prone, and inter-/intra-observer variability introduces noise to labels. This is particularly a problem when using deep learning models for segmenting medical images due to the ambiguous anatomical boundaries. Image-based medical diagnosis tools using deep learning models trained with incorrect segmentation labels can lead to false diagnoses and treatment suggestions. Multi-rater annotations might be better suited to train deep learning models with small training sets compared to single-rater annotations. The aim of this paper was to develop and evaluate a method to generate probabilistic labels based on multi-rater annotations and anatomical knowledge of the lesion features in MRI and a method to train segmentation models using probabilistic labels using normalized active-passive loss as a "noise-tolerant loss" function. The model was evaluated by comparing it to binary ground truth for 17 knees MRI scans for clinical segmentation and detection of bone marrow lesions (BML). The proposed method successfully improved precision 14, recall 22, and Dice score 8 percent compared to a binary crossentropy loss function. Overall, the results of this work suggest that the proposed normalized active-passive loss using soft labels successfully mitigated the effects of noisy labels.

show abstract

Section: ) Exploration Of Training Of An Improved Version Ofmentioning

confidence: 99%

Section: B Modelmentioning

confidence: 99%

Section: A Data and Labelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Weakly Supervised Medical Image Segmentation With Soft Labels and Noise Robust Loss

Felfeliyan¹,

Hareendranathan²,

Kuntze³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another alternative relies on a preliminary detection of the object of interest. Among the proposed tools, Mask R‐CNN, 36 an instance segmentation framework, can be considered as the state‐of‐the‐art for medical image segmentation tasks and has shown promising results in musculoskeletal 37 and breast cancer segmentation 38 …”

Section: Introductionmentioning

confidence: 99%

CNN‐based fully automatic wrist cartilage volume quantification in MR images: A comparative analysis between different CNN architectures

Vladimirov

Brui

Levchuk

et al. 2023

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose: Automatic measurement of wrist cartilage volume in MR images. Methods:We assessed the performance of four manually optimized variants of the U-Net architecture, nnU-Net and Mask R-CNN frameworks for the segmentation of wrist cartilage. The results were compared to those from a patch-based convolutional neural network (CNN) we previously designed. The segmentation quality was assessed on the basis of a comparative analysis with manual segmentation. The best networks were compared using a cross-validation approach on a dataset of 33 3D VIBE images of mostly healthy volunteers. Influence of some image parameters on the segmentation reproducibility was assessed. Results: The U-Net-based networks outperformed the patch-based CNN in terms of segmentation homogeneity and quality, while Mask R-CNN did not show an acceptable performance. The median 3D DSC value computed with the U-Net_AL (0.817) was significantly larger than DSC values computed with the other networks. In addition, the U-Net_AL provided the lowest mean volume error (17%) and the highest Pearson correlation coefficient (0.765) with respect to the ground truth values. Of interest, the reproducibility computed using U-Net_AL was larger than the reproducibility of the manual segmentation. Moreover, the results indicate that the MRI-based wrist cartilage volume is strongly affected by the image resolution.Conclusions: U-Net CNN with attention layers provided the best wrist cartilage segmentation performance. In order to be used in clinical conditions, the trained network can be fine-tuned on a dataset representing a group of specific patients.The error of cartilage volume measurement should be assessed independently using a non-MRI method.

show abstract

Weakly Supervised Medical Image Segmentation with Soft Labels and Noise Robust Loss

Felfeliyan

Hareendranathan

Kuntze

et al. 2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Improved-Mask R-CNN: Towards an accurate generic MSK MRI instance segmentation platform (data from the Osteoarthritis Initiative)

Cited by 17 publications

References 30 publications

Weakly Supervised Medical Image Segmentation With Soft Labels and Noise Robust Loss

Weakly Supervised Medical Image Segmentation With Soft Labels and Noise Robust Loss

CNN‐based fully automatic wrist cartilage volume quantification in MR images: A comparative analysis between different CNN architectures

Weakly Supervised Medical Image Segmentation with Soft Labels and Noise Robust Loss

Contact Info

Product

Resources

About