Scheduling Techniques for Liver Segmentation: ReduceLRonPlateau vs OneCycleLR

Al‐Kababji, Ayman; Bensaali, Fayçal; Dakua, Sarada Prasad

doi:10.1007/978-3-031-08277-1_17

Cited by 19 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is a result of ML algorithms' influence on more recent studies. It is quite rare to find a study that would try to compete on all the aforementioned metrics such as our study in Al-Kababji et al (2022). Moreover, most studies also refrain from using the Specificity metric, which is quite useless in the context of liver segmentation, especially when having a large volume where the liver's voxels constitute a very small part of it.…”

Section: Discussionmentioning

confidence: 99%

Automated liver tissues delineation techniques: A systematic survey on machine learning current trends and future orientations

Al‐Kababji

Bensaali

Dakua

et al. 2023

Engineering Applications of Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Automated liver tissues delineation techniques: A systematic survey on machine learning current trends and future orientations

Al‐Kababji

Bensaali

Dakua

et al. 2023

Engineering Applications of Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

“…Third, we use a learning rate scheduler, OneCycleLR, consisting of 20 epochs spent increasing the lr like a warmup and following 80 epochs with the cosine decay. It can prevent a trap at a local minimum in the training process [26]. Fourth, the parameter weight decay and dropout before the final FC layer was additionally applied to inhibit overfitting in training with a λ value of 0.01 and a dropout probability of 0.5 [27,28].…”

Section: Other Training Tricksmentioning

confidence: 99%

Enhancing Multi-disease Diagnosis of Chest X-rays with Advanced Deep-learning Networks in Real-world Data

Chen¹,

Wan

Pan

2023

J Digit Imaging

View full text Add to dashboard Cite

The current artificial intelligence (AI) models are still insufficient in multi-disease diagnosis for real-world data, which always present a long-tail distribution. To tackle this issue, a long-tail public dataset, “ChestX-ray14,” which involved fourteen (14) disease labels, was randomly divided into the train, validation, and test sets with ratios of 0.7, 0.1, and 0.2. Two pretrained state-of-the-art networks, EfficientNet-b5 and CoAtNet-0-rw, were chosen as the backbones. After the fully-connected layer, a final layer of 14 sigmoid activation units was added to output each disease’s diagnosis. To achieve better adaptive learning, a novel loss ( L ours ) was designed, which coalesced reweighting and tail sample focus. For comparison, a pretrained ResNet50 network with weighted binary cross-entropy loss ( L WBCE ) was used as a baseline, which showed the best performance in a previous study. The overall and individual areas under the receiver operating curve (AUROC) for each disease label were evaluated and compared among different models. Group-score-weighted class activation mapping (Group-CAM) is applied for visual interpretations. As a result, the pretrained CoAtNet-0-rw + L ours showed the best overall AUROC of 0.842, significantly higher than ResNet50 + L WBCE (AUROC: 0.811, p = 0.037). Group-CAM presented that the model could pay the proper attention to lesions for most disease labels (e.g., atelectasis, edema, effusion) but wrong attention for the other labels, such as pneumothorax; meanwhile, mislabeling of the dataset was found. Overall, this study presented an advanced AI diagnostic model achieving a significant improvement in the multi-disease diagnosis of chest X-rays, particularly in real-world data with challenging long-tail distributions.

show abstract

“…Both the regression and the classification models used the Adaptive Moment Estimation (Adam) optimization algorithm. Furthermore, ReduceLROnPlateauto was used to increase the learning speed and avoid overfitting [2]. Both the classification and regression model start with a 0.001 learning rate and reduce by half when the validation loss does not improve after ten epochs until the learning rate reaches 0.0001.…”

Section: Implant Size and Position Prediction Phasementioning

confidence: 99%

Automatic Placement of Simulated Dental Implants Within CBCT Images in Optimum Positions: A Deep Learning Model

Alotaibi

Alsomali

Alghamdi

et al. 2023

Preprint

View full text Add to dashboard Cite

The increasing use of three-dimensional scans in dentistry motivates the use of artificial intelligence systems in several clinical scopes. One of these scopes is implant dentistry, which is a complex process with multiple criteria of success. The dentist analyzes cone-beam computed tomography (CBCT) images and formulates a detailed implant treatment plan by specifying the length, diameter, position, and angulation of implants while considering the prosthodontic treatment plan, bone morphology, and position of adjacent vital anatomical structures. A supervised deep learning model was built to determine the optimum implant position, angulation, diameter, and length as dictated by prosthetic and anatomical requirements. The main goal of building this model was to facilitate the planning procedure and reduce the planning time of implant treatment, especially in multiple implant cases. This study had two phases; the first was detecting and placing a bounding box around the fiducial markers and adjacent bone within the CBCT images by training YOLOv3 model. The second phase trained six deep regression models to extract the apical and occlusal coordinates (X, Y, Z) of the implants; each model predicted one target value. In addition, two deep classification models were developed for predicting the implants’ intra-osseous length and intra-osseous diameter. After testing these models, YOLOv3 in the first phase showed 46.78\% average precision; in the regression model, the mean absolute error ranged from 11.931 to 15.954, and the classification of the intra-osseous diameter showed 76\% accuracy, and the intra-osseous length showed 59\% accuracy. A subjective assessment by an expert prosthodontist found the results to be promising regarding selection of implant size, and mesiodistal placement of the implants. However, angulation and vertical and bucco-lingual position of the implants was unacceptable. Thus, further investigations and model training are required to improve the accuracy of the output.

show abstract

Scheduling Techniques for Liver Segmentation: ReduceLRonPlateau vs OneCycleLR

Cited by 19 publications

References 11 publications

Automated liver tissues delineation techniques: A systematic survey on machine learning current trends and future orientations

Automated liver tissues delineation techniques: A systematic survey on machine learning current trends and future orientations

Enhancing Multi-disease Diagnosis of Chest X-rays with Advanced Deep-learning Networks in Real-world Data

Automatic Placement of Simulated Dental Implants Within CBCT Images in Optimum Positions: A Deep Learning Model

Contact Info

Product

Resources

About