Grading Framework Research Articles

A Framework for Developing Biodiversity Conservation Networks Based on Morphological Spatial Pattern Analysis and the Maximum Entropy Model: A Case Study of the Jianghan Plain, China

Constructing ecological networks in urban areas improves ecosystem stability and biodiversity protection. However, most studies focus on optimizing ecological environments through objective assessments, often neglecting species diversity. This study developed a biodiversity grading framework for the Jianghan Plain using species observation and ecosystem diversity data. Supported by ArcGIS, ecological sources were identified via MSPA and graded using the Guidelines and MaxEnt model. The MCR model was used to simulate connectivity barriers between ecological sources and calculate the minimum cumulative resistance distance, thereby generating corridors and ultimately constructing a hierarchical biodiversity conservation network for the Jianghan Plain. Our findings indicated the following: (1) The Jianghan Plain hosts 21 major ecological sources, primarily natural water bodies at the plain’s edge, which can be classified into five primary and 16 secondary sources based on biodiversity grades. (2) The recessive corridors, comprising 10 primary and 95 secondary ones, are mainly concentrated in the central Jianghan Plain, with primary corridors located centrally and westward, characterized by a large overall span. (3) Changhu Lake and Honghu Lake, two critical water bodies with high-quality habitats and significant biodiversity, were identified as key ecological nodes from the ecological sources, bridging and guiding the central and southern corridors. (4) Based on the ecological network distribution and key nodes and corridors, a “three zones, three belts, and two points” strategy was proposed for optimizing the Jianghan Plain’s ecological network. This study provides a novel framework and theoretical support for regional habitat, biodiversity conservation, and sustainable development.

HFF-Net: A hybrid convolutional neural network for diabetic retinopathy screening and grading

Diabetic Retinopathy (DR) is a leading cause of vision loss among diabetic patients, necessitating effective screening and grading for timely intervention. Regular screening significantly increases the workload of ophthalmologists, and accurate grading into stages—mild, moderate, severe, and proliferative—is crucial for monitoring disease progression. While Computer-Aided Diagnosis (CAD) systems can alleviate this burden, existing Convolutional Neural Network (CNN)-based frameworks use fixed-size kernels in a linear feed-forward manner. This approach can lead to information loss in the initial stages due to limited feature utilization across adjacent layers. To address this limitation, we propose a Hierarchical Features Fusion Convolutional Neural Network (HFF-Net) within a Diabetic Retinopathy Screening and Grading (DRSG) framework. The framework includes preprocessing to extract regions of interest from fundus images (FIs), enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE), and data augmentation for class balancing and overfitting mitigation. HFF-Net extracts multiscale features that fused at multiple levels within the network, utilizing the swish activation function for improved learning stability. We evaluated HFF-Net against several state-of-the-art CNN classifiers within the DRSG framework. Experimental results demonstrate that HFF-Net achieves a grading accuracy of 73.77 ​%, surpassing the second-best model by 3.51 percentage points (a relative improvement of approximately 5 ​%), and attains a screening accuracy of 98.70 ​% using only 1.18 million parameters. These findings highlight HFF-Net's potential as an efficient and effective tool in CAD systems for DR screening and grading.

Enhancing semi‐supervised contrastive learning through saliency map for diabetic retinopathy grading

AbstractDiabetic retinopathy (DR) is a severe ophthalmic condition that can lead to blindness if not diagnosed and provided timely treatment. Hence, the development of efficient automated DR grading systems is crucial for early screening and treatment. Although progress has been made in DR detection using deep learning techniques, these methods still face challenges in handling the complexity of DR lesion characteristics and the nuances in grading criteria. Moreover, the performance of these algorithms is hampered by the scarcity of large‐scale, high‐quality annotated data. An innovative semi‐supervised fundus image DR grading framework is proposed, employing a saliency estimation map to bolster the model's perception of fundus structures, thereby improving the differentiation between lesions and healthy regions. By integrating semi‐supervised and contrastive learning, the model's ability to recognise inter‐class and intra‐class variations in DR grading is enhanced, allowing for precise discrimination of various lesion features. Experiments conducted on publicly available DR grading datasets, such as EyePACS and Messidor, have validated the effectiveness of our proposed method. Specifically, our approach outperforms the state of the art on the kappa metric by 0.8% on the full EyePACS dataset and by 3.2% on a 10% subset of EyePACS, demonstrating its superiority over previous methodologies. The authors’ code is publicly available at https://github.com/500ZhangJC/SCL‐SEM‐framework‐for‐DR‐Grading.

Nakagami-fuzzy imaging for grading brain tumors by analyzing fractal complexity

Gliomas are the brain tumors in glial cells, which are categorized into four numerical grades, I-II-III-IV, to quantize the aggressiveness and severity of the tumors; while divided into two major groups, high-grade (HG) and low-grade (LG), in general. Among many differences between these groups, one of the most distinct and characteristic features could be seen in the shape of the tumor boundaries by magnetic resonance imaging (MRI). Due to aggressive nature of the HG tumors in proliferation phase, the boundaries of HG tumors become more shape-wise complex compared to the LG tumors, which could be differentiated by analyzing the fractal complexity of the cell membranes. However, the complexity cannot be either manually calculated or estimated by eye inspection without a reference point with one single image or sometimes even with an image set. Therefore, we present an automated glioma grading framework to provide an insight on the grades with a novel contouring and fractal dimension analysis system. The primary component of the proposed system is an automated Nakagami imaging module with a specialized fuzzy c-means algorithm to contour the boundaries of the whole tumors. The contoured images, afterwards, are analyzed by the Minkowski–Bouligand and Hausdorff methods for two panning options to generate the fractal dimensions and to estimate the fractal complexities for classifying the gliomas The results are greatly encouraging that the overall classification accuracy is computed as 88.31 % using the basic support vector machines (SVM) classifier; while as 91.96 % with the arbitrary thresholding appended. The outcomes of this paper with implementable mathematical infrastructure would be very useful and beneficial as an expert system in intelligent and automatic glioma grading, for researchers and medical experts.

Robust, credible, and interpretable AI-based histopathological prostate cancer grading.

Prostate cancer (PCa) is among the most common cancers in men and its diagnosis requires the histopathological evaluation of biopsies by human experts. While several recent artificial intelligence-based (AI) approaches have reached human expert-level PCa grading, they often display significantly reduced performance on external datasets. This reduced performance can be caused by variations in sample preparation, for instance the staining protocol, section thickness, or scanner used. Another limiting factor of contemporary AI-based PCa grading is the prediction of ISUP grades, which leads to the perpetuation of human annotation errors. We developed the prostate cancer aggressiveness index (PCAI), an AI-based PCa detection and grading framework that is trained on objective patient outcome, rather than subjective ISUP grades. We designed PCAI as a clinical application, containing algorithmic modules that offer robustness to data variation, medical interpretability, and a measure of prediction confidence. To train and evaluate PCAI, we generated a multicentric, retrospective, observational trial consisting of six cohorts with 25,591 patients, 83,864 images, and 5 years of median follow-up from 5 different centers and 3 countries. This includes a high-variance dataset of 8,157 patients and 28,236 images with variations in sample thickness, staining protocol, and scanner, allowing for the systematic evaluation and optimization of model robustness to data variation. The performance of PCAI was assessed on three external test cohorts from two countries, comprising 2,255 patients and 9,437 images. Using our high-variance datasets, we show how differences in sample processing, particularly slide thickness and staining time, significantly reduce the performance of AI-based PCa grading by up to 6.2 percentage points in the concordance index (C-index). We show how a select set of algorithmic improvements, including domain adversarial training, conferred robustness to data variation, interpretability, and a measure of credibility to PCAI. These changes lead to significant prediction improvement across two biopsy cohorts and one TMA cohort, systematically exceeding expert ISUP grading in C-index and AUROC by up to 22 percentage points. Data variation poses serious risks for AI-based histopathological PCa grading, even when models are trained on large datasets. Algorithmic improvements for model robustness, interpretability, credibility, and training on high-variance data as well as outcome-based severity prediction gives rise to robust models with above ISUP-level PCa grading performance.

CauDR: A causality-inspired domain generalization framework for fundus-based diabetic retinopathy grading

Diabetic retinopathy (DR) is the most common diabetic complication, which usually leads to retinal damage, vision loss, and even blindness. A computer-aided DR grading system has a significant impact on helping ophthalmologists with rapid screening and diagnosis. Recent advances in fundus photography have precipitated the development of novel retinal imaging cameras and their subsequent implementation in clinical practice. However, most deep learning-based algorithms for DR grading demonstrate limited generalization across domains. This inferior performance stems from variance in imaging protocols and devices inducing domain shifts. We posit that declining model performance between domains arises from learning spurious correlations in the data. Incorporating do-operations from causality analysis into model architectures may mitigate this issue and improve generalizability. Specifically, a novel universal structural causal model (SCM) was proposed to analyze spurious correlations in fundus imaging. Building on this, a causality-inspired diabetic retinopathy grading framework named CauDR was developed to eliminate spurious correlations and achieve more generalizable DR diagnostics. Furthermore, existing datasets were reorganized into 4DR benchmark for DG scenario. Results demonstrate the effectiveness and the state-of-the-art (SOTA) performance of CauDR. Diabetic retinopathy (DR) is the most common diabetic complication, which usually leads to retinal damage, vision loss, and even blindness. A computer-aided DR grading system has a significant impact on helping ophthalmologists with rapid screening and diagnosis. Recent advances in fundus photography have precipitated the development of novel retinal imaging cameras and their subsequent implementation in clinical practice. However, most deep learning-based algorithms for DR grading demonstrate limited generalization across domains. This inferior performance stems from variance in imaging protocols and devices inducing domain shifts. We posit that declining model performance between domains arises from learning spurious correlations in the data. Incorporating do-operations from causality analysis into model architectures may mitigate this issue and improve generalizability. Specifically, a novel universal structural causal model (SCM) was proposed to analyze spurious correlations in fundus imaging. Building on this, a causality-inspired diabetic retinopathy grading framework named CauDR was developed to eliminate spurious correlations and achieve more generalizable DR diagnostics. Furthermore, existing datasets were reorganized into 4DR benchmark for DG scenario. Results demonstrate the effectiveness and the state-of-the-art (SOTA) performance of CauDR.

Development and validation of the Sinonasal Endoscopic Score (SiNES) for chronic rhinosinusitis.

Although there are several endoscopic grading systems for chronic rhinosinusitis (CRS), they are limited in their range and applicability. We developed a SiNonasal Endoscopic Score (SiNES) that builds upon the strengths of previous systems while addressing their limitations. The SiNES system was developed by consensus after multiple rounds of guided discussions. Face, content, and convergent validity were investigated. It was validated using an independent sample of 79 CRS individuals from two referral centres from September 2021 to February 2022. Each patient underwent a sinonasal endoscopy and filled PROM questionnaires. Three independent rhinologists graded endoscopic videos using the SiNES and modified Lund-Kennedy (MLK) scores. Inter-rater and test-retest reliability were assessed via the intraclass correlation coefficient (ICC). SiNES and MLK scores were correlated with PROMs using a Spearman correlation and canonical correlation analysis (CCA). The SiNES system evaluates five anatomical spaces regarding edema, discharge, and scarring. Face, content, and convergent validity were deemed satisfactory by the study authors and an independent panel of Otolaryngologists. Inter-rater reliability was excellent for the SiNES and good for the MLK score. Test-retest reliability was excellent for both systems. Total SiNES was correlated with self-reported smell loss. The SiNES system is an accurate and reliable grading framework applicable to all CRS subtypes. It can be utilized in clinical and research settings and improves upon previously published systems.

Life cycle and techno-economic assessment of bioresource production from wastewater

Thermochemical conversion technologies present an opportunity to flip the paradigm of wastewater biosolids management operations from energy-intense and expensive waste management processes into energy-positive and economical resource extraction centers. Herein, we present a uniform “grading framework” to consistently evaluate the environmental and commercial benefits of established and emerging wastewater biosolids management processes from a life cycle and techno-economic perspective. Application of this approach reveals that established wastewater biosolids management practices such as landfilling, land application, incineration, and anaerobic digestion, while commercially viable, offer little environmental benefit. On the other hand, emerging thermochemical bioresource recovery technologies such as hydrothermal liquefaction, gasification, pyrolysis, and torrefaction show potential to provide substantial economic and environmental benefit through the recovery of carbon and nutrients from wastewater biosolids in the form of biofuels, fertilizers, and other high-value products. Some emerging thermochemical technologies have developed beyond pilot scale although their commercial viability remains to be seen.

Prostate cancer grading framework based on deep transfer learning and Aquila optimizer

Prostate cancer is the one of the most dominant cancer among males. It represents one of the leading cancer death causes worldwide. Due to the current evolution of artificial intelligence in medical imaging, deep learning has been successfully applied in diseases diagnosis. However, most of the recent studies in prostate cancer classification suffers from either low accuracy or lack of data. Therefore, the present work introduces a hybrid framework for early and accurate classification and segmentation of prostate cancer using deep learning. The proposed framework consists of two stages, namely classification stage and segmentation stage. In the classification stage, 8 pretrained convolutional neural networks were fine-tuned using Aquila optimizer and used to classify patients of prostate cancer from normal ones. If the patient is diagnosed with prostate cancer, segmenting the cancerous spot from the overall image using U-Net can help in accurate diagnosis, and here comes the importance of the segmentation stage. The proposed framework is trained on 3 different datasets in order to generalize the framework. The best reported classification accuracies of the proposed framework are 88.91% using MobileNet for the “ISUP Grade-wise Prostate Cancer” dataset and 100% using MobileNet and ResNet152 for the “Transverse Plane Prostate Dataset” dataset with precisions 89.22% and 100%, respectively. U-Net model gives an average segmentation accuracy and AUC of 98.46% and 0.9778, respectively, using the “PANDA: Resized Train Data (512 × 512)” dataset. The results give an indicator of the acceptable performance of the proposed framework.

A fusion deep learning framework based on breast cancer grade prediction

In breast cancer grading, the subtle differences between HE-stained pathological images and the insufficient number of data samples lead to grading inefficiency. With its rapid development, deep learning technology has been widely used for automatic breast cancer grading based on pathological images. In this paper, we propose an integrated breast cancer grading framework based on a fusion deep learning model, which uses three different convolutional neural networks as submodels to extract feature information at different levels from pathological images. Then, the output features of each submodel are learned by the fusion network based on stacking to generate the final decision results. To validate the effectiveness and reliability of our proposed model, we perform dichotomous and multiclassification experiments on the Invasive Ductal Carcinoma (IDC) pathological image dataset and a generated dataset and compare its performance with those of the state-of-the-art models. The classification accuracy of the proposed fusion network is 93.8%, the recall is 93.5%, and the F1 score is 93.8%, which outperforms the state-of-the-art methods.

Journey to an “A” Grade Gas Certification

The August 2022 issue of JPT featured an article on the most common voluntary certifications that have at least some kind of methane emissions-related requirements. Since then, MiQ appears to have become a certification framework of choice for many industry players. According to Georges Tijbosch, CEO of MiQ, as of January 2023, about 20% of US natural gas production has gone through the MiQ process. Tijbosch noted that most of the currently certified production received grades A, B, or C. While this is no high school, we naturally tend to assume that getting an “A” is the ultimate goal of getting certified. It might be true for some; however, as we all realize after graduating schools and universities and upon entering the workforce, life is quite often more about continuous improvement over time vs. achieving the top grade as such. This article is aimed at exploring the MiQ grading framework in more detail and explaining practical pathways for industry participants to achieve any desired MiQ grade. For simplicity, we will focus on the MiQ Standard—Onshore Production, but the fundamental concepts are applicable to other industry segments. The MiQ Standard scores a facility based on three broad categories. - Methane-intensity quantitative metric - Monitoring technology deployment - Company practices related to the methane emissions management Each of these categories is scored by an auditor accredited to conduct MiQ audits and who is an independent subject matter expert in the fields of methane mitigation, monitoring, and accounting. The combined score across the three categories results in a letter grade on the MiQ certificate issued for that facility’s production. Natural gas buyers (typically utility companies or industrial users) then request certain volumes of certified gas and, depending on the targeted end use and other factors, may specify a desired grade. Roy Hartstein is an emissions subject matter expert, as well as founder and president of Responsible Energy Solutions, an organization that performs certification assessments and helps companies prepare for certification. Hartstein and his team have conducted over a dozen MiQ audits. He talks about three aspects of value that motivate natural gas producers to improve from a non-A to an A grade: 1) possibility of earning a higher premium on the sold certified gas; 2) internal risk management across several avenues such as safety, identification of malfunctioning equipment, or early detection of operational inefficiencies; and 3) alleviating environmental concerns of the investment community. He believes this advances a long-term strategy to secure oil and gas’s potential as part of the future energy mix. Regardless of the underlying motivation, most of us like to think we are straight-A students and don’t want to bring home a report card with a lower grade. But at the time a company initiates the certification process, they may find they qualify for a lower (i.e., non-A) grade. What are the potential implications of this?

Peer Grading Eliciting Truthfulness Based on Autograder

Peer grading has diverse applications in many fields, including the peer grading of open assignments in online courses. The major challenge in peer grading is improving the seriousness (review carefully) of reviewers. The previous studies have proposed a number of incentive reward mechanisms intended to reward or punish reviewers. Although these mechanisms are effective in massive open online courses environments with a large number of students, they are not suitable for small private online courses environments with a small number of students. This study analysis the characteristics of reviewers in a small private online course environment and designs an incentive and reward mechanism based on a hybrid human-machine peer grading framework, including an auto-grader (reviewed by the machine). The main component of the proposed mechanism is the auto-grader whose function is to evaluate the reviewers' reports (reviewers refer to students) and provide feedback information on reports to reviewers. The framework can incentivize reviewers to review carefully and give their reports truthfully when the reports between reviewers' and the auto-grader' are consistent. The proposed framework is verified experimentally. The experimental results demonstrate that the hybrid human-machine peer grading framework can effectively incentivize reviewers to review submissions carefully and improve the overall effect of peer grading.

A multimodal AI-based non-invasive COVID-19 grading framework powered by deep learning, manta ray, and fuzzy inference system from multimedia vital signs

The COVID-19 pandemic has presented unprecedented challenges to healthcare systems worldwide. One of the key challenges in controlling and managing the pandemic is accurate and rapid diagnosis of COVID-19 cases. Traditional diagnostic methods such as RT-PCR tests are time-consuming and require specialized equipment and trained personnel. Computer-aided diagnosis systems and artificial intelligence (AI) have emerged as promising tools for developing cost-effective and accurate diagnostic approaches. Most studies in this area have focused on diagnosing COVID-19 based on a single modality, such as chest X-rays or cough sounds. However, relying on a single modality may not accurately detect the virus, especially in its early stages. In this research, we propose a non-invasive diagnostic framework consisting of four cascaded layers that work together to accurately detect COVID-19 in patients. The first layer of the framework performs basic diagnostics such as patient temperature, blood oxygen level, and breathing profile, providing initial insights into the patient's condition. The second layer analyzes the coughing profile, while the third layer evaluates chest imaging data such as X-ray and CT scans. Finally, the fourth layer utilizes a fuzzy logic inference system based on the previous three layers to generate a reliable and accurate diagnosis. To evaluate the effectiveness of the proposed framework, we used two datasets: the Cough Dataset and the COVID-19 Radiography Database. The experimental results demonstrate that the proposed framework is effective and trustworthy in terms of accuracy, precision, sensitivity, specificity, F1-score, and balanced accuracy. The audio-based classification achieved an accuracy of 96.55%, while the CXR-based classification achieved an accuracy of 98.55%. The proposed framework has the potential to significantly improve the accuracy and speed of COVID-19 diagnosis, allowing for more effective control and management of the pandemic. Furthermore, the framework's non-invasive nature makes it a more attractive option for patients, reducing the risk of infection and discomfort associated with traditional diagnostic methods.

Identifying the Key Components in ResNet-50 for Diabetic Retinopathy Grading from Fundus Images: A Systematic Investigation.

Although deep learning-based diabetic retinopathy (DR) classification methods typically benefit from well-designed architectures of convolutional neural networks, the training setting also has a non-negligible impact on prediction performance. The training setting includes various interdependent components, such as an objective function, a data sampling strategy, and a data augmentation approach. To identify the key components in a standard deep learning framework (ResNet-50) for DR grading, we systematically analyze the impact of several major components. Extensive experiments are conducted on a publicly available dataset EyePACS. We demonstrate that (1) the DR grading framework is sensitive to input resolution, objective function, and composition of data augmentation; (2) using mean square error as the loss function can effectively improve the performance with respect to a task-specific evaluation metric, namely the quadratically weighted Kappa; (3) utilizing eye pairs boosts the performance of DR grading and; (4) using data resampling to address the problem of imbalanced data distribution in EyePACS hurts the performance. Based on these observations and an optimal combination of the investigated components, our framework, without any specialized network design, achieves a state-of-the-art result (0.8631 for Kappa) on the EyePACS test set (a total of 42,670 fundus images) with only image-level labels. We also examine the proposed training practices on other fundus datasets and other network architectures to evaluate their generalizability. Our codes and pre-trained model are available online.

MCLSG:Multi-modal classification of lung disease and severity grading framework using consolidated feature engineering mechanisms

Human disease detection using medical images with algorithmic severity prediction is unexplored. Deep learning algorithms for disease identification and classification from medical images have been proposed recently. Precision medicine requires more than illness detection. Expert diagnosis and advanced medical examinations are frequently necessary to decrease the likelihood of cross-infection and cater to a patient based on his severity level. To enable the assessment of human disease detection with its severity, we propose the Multi-modal Classification of Lung Disease and Severity Grading (MCLSG) framework from the different medical image modalities using dedicated feature engineering mechanisms. The automatic MCLSG framework consists of pre-processing, automatic feature extraction, classification, and severity analysis. The goal of pre-processing phase is to enhance the image quality and localize the region of interest (ROI) without using any segmentation method. The unwanted image regions are cropped followed by the quality enhancement in the pre-processing. In the automatic feature extraction phase, the optimized Convolutional Neural Network (CNN) layers perform the automatic feature learning and selection using pre-trained parameters. Optimized CNN is then combined with machine-learning classifiers for disease identification. When the disease is detected, MCLSG does automatic disease severity assessments utilizing a consolidated feature engineering technique that includes high-level and low-level characteristics rather than only CNN features. According to the outcome of consolidated feature engineering, we divide patients into four severity levels (mild, moderate, severe, and fatal). The experimental findings show that the proposed MCLSG model is efficient. Overall results show that MCLSG improves the accuracy by 3.5 % for disease detection and 6.8 % for severity level analysis.

MsgeCNN: Multiscale geometric embedded convolutional neural network for ONFH segmentation and grading.

The incidence of osteonecrosis of the femoral head (ONFH) is increasing gradually, rapid and accurate grading of ONFH is critical. The existing Steinberg staging criteria grades ONFH according to the proportion of necrosis area to femoral head area. In the clinical practice, the necrosis region and femoral head region are mainly estimated by the observation and experience of doctor. This paper proposes a two-stage segmentation and grading framework, which can be used to segment the femoral head and necrosis, as well as todiagnosis. The core of the proposed two-stage framework is the multiscale geometric embedded convolutional neural network (MsgeCNN), which integrates geometric information into the training process and accurately segments the femoral head region. Then, the necrosis regions are segmented by the adaptive threshold method taking femoral head as the background. The area and proportion of the two are calculated to determine thegrade. The accuracy of the proposed MsgeCNN for femoral head segmentation is 97.73%, sensitivity is 91.17%, specificity is 99.40%, dice score is 93.34%. And the segmentation performance is better than the existing five segmentation algorithms. The diagnostic accuracy of the overall framework is 90.80%. The proposed framework can accurately segment the femoral head region and the necrosis region. The area, proportion, and other pathological information of the framework output provide auxiliary strategies for subsequent clinicaltreatment.

Fruit maturity grading framework for small dataset using single image multi-object sampling and Mask R-CNN

Fruit maturity grading is the key factor for fruit export where maturity consistence and standard are required. This paper proposes a non-destructive approach for pineapple maturity grading and pineapple localization based on object segmentation framework which has enhanced for training a robust model with small dataset. We introduced a multi-object sampling technique in augmentation process to generate images from a small dataset taken under controlled conditions to generalize the model for a more practical dataset. We identify the robustness of object segmentation models, Mask R-CNN over other models, e.g., Faster R-CNN, RetinaNet and CenterMask through mean average precision (mAP), detection ratio to explore the false positives detected, precision-recall curve and computational time. The optimal threshold selection which is crucial especially for sensitive small dataset to achieve high detection performance is proposed in this work using mAP and detection ratio. Our proposed framework enhances the model generalization and achieves mAP of 86.7%, AP50 and AP75 at 97.98% and APunripe, APpartially_ripe and APfully_ripe of 99.20%, 96.58% and 98.63%, respectively. Additional insights of the models developed with our small dataset are explained along with the experimental results which are suggested for future.

KIEGLFN: A unified acne grading framework on face images

Background and Objective: Grading the severity level is an extremely important procedure for correct diagnoses and personalized treatment schemes for acne. However, the acne grading criteria are not unified in the medical field. This work aims to develop an acne diagnosis system that can be generalized to various criteria. Methods: A unified acne grading framework that can be generalized to apply referring to different grading criteria is developed. It imitates the global estimation of the dermatologist diagnosis in two steps. First, an adaptive image preprocessing method effectively filters meaningless information and enhances key information. Next, an innovative network structure fuses global deep features with local features to simulate the dermatologists’ comparison of local skin and global observation. In addition, a transfer fine-tuning strategy is proposed to transfer prior knowledge on one criterion to another criterion, which effectively improves the framework performance in case of insufficient data. Results: The Preprocessing method effectively filters meaningless areas and improves the performance of downstream models.The framework reaches accuracies of 84.52% and 59.35% on two datasets separately. Conclusions: The application of the framework on acne grading exceeds the state-of-the-art method by 1.71%, reaches the diagnostic level of a professional dermatologist and the transfer fine-tuning strategy improves the accuracy of 6.5% on the small data.

Using the Dynamic Shear Rheometer for Low-Temperature Grading of Asphalt Binders

Abstract The Superpave Performance Grading (PG) framework or a variation of this framework is often used as a purchase specification in the United States and as a tool to evaluate the expected performance of modified and unmodified asphalt binders in many countries around the world. A cornerstone of the PG framework is the low-temperature grade of the asphalt binder that is assessed using a bending beam rheometer (BBR) using maximum stiffness and a minimum rate of deformation as criteria. Several previous studies from other researchers have demonstrated the feasibility of using the dynamic shear rheometer (DSR) with a 4-mm parallel plate geometry to measure the rheological properties of asphalt binders at low temperatures. This article builds on these studies to evaluate the feasibility of using a DSR to replace the use of a BBR for low-temperature grading of the asphalt binder using two different approaches. The first approach was to measure complex shear modulus of different asphalt binders using the DSR, convert these measurements to BBR equivalent stiffness and m-value using principles of linear viscoelasticity, and compare the two pairs of parameters from each method. The second approach was to use the complex shear modulus and phase angle measured using the DSR in lieu of using the stiffness and m-value measured using the BBR. Although the results from both methods show feasibility of using the 4-mm plate with the DSR to estimate low-temperature properties, the second approach of measuring and using the DSR-based complex shear moduli and phase angle is a more viable alternative for routine use.

Deep Learning-Based Automatic Assessment of Radiation Dermatitis in Patients With Nasopharyngeal Carcinoma

Radiation dermatitis (RD) is a common, unpleasant side effect of patients receiving radiation therapy. In clinical practice, the severity of RD is graded manually through visual inspection, which is labor intensive and often leads to large interrater variations. To overcome these shortcomings, this study aimed to develop an automatic RD assessment based on deep learning (DL) techniques that could efficiently assist the RD severity classification in clinical application. A total of 1205 photographs of the head and neck region were collected from patients with nasopharyngeal carcinoma (NPC) undergoing radiation therapy. The severity of RD in these photographs was graded by 5 qualified assessors based on the Radiation Therapy Oncology Group guidance. An end-to-end RD grading framework was developed by combining a DL-based segmentation network and a DL-based RD severity classifier, which are used for segmenting the neck region from the camera-captured photographs and grading, respectively. U-Net was used for segmentation and another convolutional neural network classifier (DenseNet-121) was applied to RD severity classification. Dice similarity coefficient was used to evaluate the performance of segmentation. Severity classification was evaluated by several metrics, including overall accuracy, precision, recall, and F1 score. Results of segmentation showed that the averaged dice similarity coefficients were 91.2% and 90.8% for front and side view, respectively. For RD severity classification, the overall accuracy of test photographs was 83.0%. Our method accurately classified 90.5% of grade 0, 67.2% of grade 1, 93.8% of grade 2, and 100% of above grade 2 cases. The overall prediction performance was comparable with human assessors. There was no significant difference in accuracy when using manually or automatically segmented regions (P=.683). We have successfully demonstrated a DL-based method for automatic assessment of RD severity in patients with NPC. This method holds great potential for efficient and effective assessing and monitoring of RD in patients with NPC.