Journal: Wearable Technologies

Article Title: A deep learning method to predict ankle joint moment during walking at different speeds with ultrasound imaging: A framework for assistive devices control

doi: 10.1017/wtc.2022.18

Figure Lengend Snippet: Data collection, pre-processing, and schematic illustration of the proposed deep CNN model calibration. The input to the CNN model includes either the time sequence data of cropped US imaging with different regions of interest or the time sequence data of sEMG spectrum imaging. Thirty-one layers were created in the designed CNN model, including one image input layer at the beginning, one fully connected layer, one regression output layer at the end, and seven sets of intermediate layers. Each intermediate set contained one convolution 2D layer, one batch normalization layer, one rectified linear unit (ReLU) layer, and one average pooling 2D layer.

Article Snippet: The designed deep CNN model with 31 layers in total was created utilizing Matlab (R2020a, MathWorks, MA).

Techniques: Sequencing, Imaging

Journal: Wearable Technologies

Article Title: A deep learning method to predict ankle joint moment during walking at different speeds with ultrasound imaging: A framework for assistive devices control

doi: 10.1017/wtc.2022.18

Figure Lengend Snippet: Ankle joint net plantarflexion moment prediction time sequence on each participant by using US images with ROI of 100 × 100 pixels and the deep learning approach. The red solid and blue dashed curves represent the measurements from inverse dynamics and prediction from the CNN model. For each walking speed, three walking stance cycles are included for prediction, therefore, 15 periodic curves are shown for each participant (with the speed order of 0.50, 0.75, 1.00, 1.25, and 1.50 m/s).

Article Snippet: The designed deep CNN model with 31 layers in total was created utilizing Matlab (R2020a, MathWorks, MA).

Techniques: Sequencing

Journal: Wearable Technologies

Article Title: A deep learning method to predict ankle joint moment during walking at different speeds with ultrasound imaging: A framework for assistive devices control

doi: 10.1017/wtc.2022.18

Figure Lengend Snippet: Ankle joint net plantarflexion moment prediction as a percentage of the stance cycle (0% for heel-strike and 100% for toe-off) by using US images with ROI of 100 × 100 pixels and the deep learning approach. The red and blue center curves and shadowed areas represent the mean and standard deviation values (three stance cycles for each curve) of the ground truth and CNN model-based prediction, respectively. Each row subplots represent data from individual participant while each column subplots represent individual walking speed out of five.

Article Snippet: The designed deep CNN model with 31 layers in total was created utilizing Matlab (R2020a, MathWorks, MA).

Techniques: Standard Deviation

Journal: Wearable Technologies

Article Title: A deep learning method to predict ankle joint moment during walking at different speeds with ultrasound imaging: A framework for assistive devices control

doi: 10.1017/wtc.2022.18

Figure Lengend Snippet: The individual net plantarflexion moment prediction RMSE and N-RMSE values of 15 stance cycles across five walking speeds by using the trained personalized CNN model with different ROIs.

Article Snippet: The designed deep CNN model with 31 layers in total was created utilizing Matlab (R2020a, MathWorks, MA).

Techniques:

Journal: Journal of Imaging

Article Title: Full-Reference Image Quality Assessment Based on an Optimal Linear Combination of Quality Measures Selected by Simulated Annealing

doi: 10.3390/jimaging8080224

Figure Lengend Snippet: PLCC, SROCC, and KROCC performance comparison of the proposed fusion-based FR-IQA metrics on LIVE and TID2013 databases with the state-of-the-art. The best results are typed in bold, and the second best results are underlined.

Article Snippet: Moreover, we reimplemented the fusion-based SSIM-CNN [ ] method in MATLAB R2020a (available at: https://github.com/Skythianos/SSIM-CNN (accessed on 12 May 2022)).

Techniques:

Journal: Journal of Imaging

Article Title: Full-Reference Image Quality Assessment Based on an Optimal Linear Combination of Quality Measures Selected by Simulated Annealing

doi: 10.3390/jimaging8080224

Figure Lengend Snippet: PLCC, SROCC, and KROCC performance comparison of the proposed fusion-based FR-IQA metrics on TID2008 and CSIQ databases with the state-of-the-art. The best results are typed in bold, and the second best results are underlined.

Article Snippet: Moreover, we reimplemented the fusion-based SSIM-CNN [ ] method in MATLAB R2020a (available at: https://github.com/Skythianos/SSIM-CNN (accessed on 12 May 2022)).

Techniques:

Journal: Journal of Imaging

Article Title: Full-Reference Image Quality Assessment Based on an Optimal Linear Combination of Quality Measures Selected by Simulated Annealing

doi: 10.3390/jimaging8080224

Figure Lengend Snippet: PLCC, SROCC, and KROCC performance comparison of the proposed fusion-based FR-IQA metrics with the state-of-the-art. The best results are typed in bold, the second best results are underlined.

Article Snippet: Moreover, we reimplemented the fusion-based SSIM-CNN [ ] method in MATLAB R2020a (available at: https://github.com/Skythianos/SSIM-CNN (accessed on 12 May 2022)).

Techniques: