Ms RED: A novel multi-scale residual encoding and decoding network for skin lesion segmentation

Dai, Duwei; Dong, Caixia; Yan, Qingsen; Li, Zongfang; Zhang, Chunyan; Luo, Nana

doi:10.1016/j.media.2021.102293

Cited by 122 publications

(57 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…( 13 ). Deeper and restructured, ResNet101 has shown a high performance in many contexts of use including skin lesion detection and brain disease detection in magnetic resonance images ( 14 , 15 ).…”

Section: Methodsmentioning

confidence: 99%

A Deep Learning System to Predict the Histopathological Results From Urine Cytopathological Images

et al. 2022

View full text Add to dashboard Cite

BackgroundAlthough deep learning systems (DLSs) have been developed to diagnose urine cytology, more evidence is required to prove if such systems can predict histopathology results as well.MethodsWe retrospectively retrieved urine cytology slides and matched histological results. High-power field panel images were annotated by a certified urological pathologist. A deep learning system was designed with a ResNet101 Faster R-CNN (faster region-based convolutional neural network). It was firstly built to spot cancer cells. Then, it was directly used to predict the likelihood of the presence of tissue malignancy.ResultsWe retrieved 441 positive cases and 395 negative cases. The development involved 387 positive cases, accounting for 2,668 labeled cells, to train the DLS to spot cancer cells. The DLS was then used to predict corresponding histopathology results. In an internal test set of 85 cases, the area under the curve (AUC) was 0.90 (95%CI 0.84–0.96), and the kappa score was 0.68 (95%CI 0.52–0.84), indicating substantial agreement. The F1 score was 0.56, sensitivity was 71% (95%CI 52%–85%), and specificity was 94% (95%CI 84%–98%). In an extra test set of 333 cases, the DLS achieved 0.25 false-positive cells per image. The AUC was 0.93 (95%CI 0.90–0.95), and the kappa score was 0.58 (95%CI 0.46–0.70) indicating moderate agreement. The F1 score was 0.66, sensitivity was 67% (95%CI 54%–78%), and specificity was 92% (95%CI 88%–95%).ConclusionsThe deep learning system could predict if there was malignancy using cytocentrifuged urine cytology images. The process was explainable since the prediction of malignancy was directly based on the abnormal cells selected by the model and can be verified by examining those candidate abnormal cells in each image. Thus, this DLS was not just a tool for pathologists in cytology diagnosis. It simultaneously provided novel histopathologic insights for urologists.

show abstract

Section: Methodsmentioning

confidence: 99%

A Deep Learning System to Predict the Histopathological Results From Urine Cytopathological Images

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Yet FAT-Net may effectively extract local features and global true label whereas CNN are not capable of learning global true labels sufficiently [ 24 ]. Similarly, a neural network-based Multi-scale Residual Encoding and Decoding network (Ms RED) is used to handle blurred boundaries [ 25 ]. Thapar et al [ 26 ] employed a segmentation framework using swarm intelligence with Grasshopper Optimization Algorithm (GOA) for feature extraction and successfully obtained 98.42% classification accuracy.…”

Section: Literature Surveymentioning

confidence: 99%

A Framework for Interactive Medical Image Segmentation Using Optimized Swarm Intelligence with Convolutional Neural Networks

Kaushal

Islam

Althubiti

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Recent improvements in current technology have had a significant impact on a wide range of image processing applications, including medical imaging. Classification, detection, and segmentation are all important aspects of medical imaging technology. An enormous need exists for the segmentation of diagnostic images, which can be applied to a wide variety of medical research applications. It is important to develop an effective segmentation technique based on deep learning algorithms for optimal identification of regions of interest and rapid segmentation. To cover this gap, a pipeline for image segmentation using traditional Convolutional Neural Network (CNN) as well as introduced Swarm Intelligence (SI) for optimal identification of the desired area has been proposed. Fuzzy C-means (FCM), K-means, and improvisation of FCM with Particle Swarm Optimization (PSO), improvisation of K-means with PSO, improvisation of FCM with CNN, and improvisation of K-means with CNN are the six modules examined and evaluated. Experiments are carried out on various types of images such as Magnetic Resonance Imaging (MRI) for brain data analysis, dermoscopic for skin, microscopic for blood leukemia, and computed tomography (CT) scan images for lungs. After combining all of the datasets, we have constructed five subsets of data, each of which had a different number of images: 50, 100, 500, 1000, and 2000. Each of the models was executed and trained on the selected subset of the datasets. From the experimental analysis, it is observed that the performance of K-means with CNN is better than others and achieved 96.45% segmentation accuracy with an average time of 9.09 seconds.

show abstract

“…Several ground truths, for the same image, were used to train several networks, which were then fused to obtain the final prediction. Other work also utilized the fusion of several image scales to improve the quality of the segmentation [ 49 ]. Lastly, the difference between image acquisition formats was also analyzed.…”

Section: Background and Related Workmentioning

confidence: 99%

Integrating Domain Knowledge into Deep Learning for Skin Lesion Risk Prioritization to Assist Teledermatology Referral

et al. 2021

View full text Add to dashboard Cite

Teledermatology has developed rapidly in recent years and is nowadays an essential tool for early diagnosis. In this work, we aim to improve existing Teledermatology processes for skin lesion diagnosis by developing a deep learning approach for risk prioritization with a dataset of retrospective data from referral requests of the Portuguese National Health System. Given the high complexity of this task, we propose a new prioritization pipeline guided and inspired by domain knowledge. We explored automatic lesion segmentation and tested different learning schemes, namely hierarchical classification and curriculum learning approaches, optionally including additional patient metadata. The final priority level prediction can then be obtained by combining predicted diagnosis and a baseline priority level accounting for explicit expert knowledge. In both the differential diagnosis and prioritization branches, lesion segmentation with 30% tolerance for contextual information was shown to improve classification when compared with a flat baseline model trained on original images; furthermore, the addition of patient information was not beneficial for most experiments. Curriculum learning delivered better results than a flat or hierarchical approach. The combination of diagnosis information and a knowledge map, created in collaboration with dermatologists, together with the priority achieved interesting results (best macro F1 of 43.93% for a validated test set), paving the way for new data-centric and knowledge-driven approaches.

show abstract

Ms RED: A novel multi-scale residual encoding and decoding network for skin lesion segmentation

Cited by 122 publications

References 25 publications

A Deep Learning System to Predict the Histopathological Results From Urine Cytopathological Images

A Deep Learning System to Predict the Histopathological Results From Urine Cytopathological Images

A Framework for Interactive Medical Image Segmentation Using Optimized Swarm Intelligence with Convolutional Neural Networks

Integrating Domain Knowledge into Deep Learning for Skin Lesion Risk Prioritization to Assist Teledermatology Referral

Contact Info

Product

Resources

About