A Deep Learning Instance Segmentation Approach for Global Glomerulosclerosis Assessment in Donor Kidney Biopsies

Extended Criteria Donor Kidney Transplantation: Comparative Outcome Analysis Between Single versus Double Kidney Transplantation at 5 Years

Structural analysis, including morphometric computation of glomerular size and number, was applied to analyse the divergence between propensity to hypertension and renal damage, expressed by rats of the Milan hypertensive strain (MHS) and Milan normotensive strain (MNS), respectively. MHS, MNS rats and progenitor Wistar rats were investigated at age 9 weeks and 9 months. Classical morphometric methods were complemented by the dissector/fractionator technique to count glomeruli. At 9 weeks, when nephrogenesis was completed and hypertension established, MHS rats exhibited significantly lower kidney weight, cortical volume, glomerular number and volume compared to coeval MNS rats. In Wistar rats, these parameters were similar to those of MNS rats, except for lower glomerular volume. At 9 months, MHS rats showed significantly lower expansion of glomerular volume compared to MNS and Wistar rats. MNS rats had 10% sclerotic glomeruli, which was associated with reduced renal function and heavy proteinuria; conversely, sclerosis was rare in coeval MHS and Wistar rats. Media thickness was higher, whereas lumen diameter was lower, in intrarenal arteries of MHS versus MNS rats at both time points. These data indicate that structural changes other than a tubular defect may play a role in the development of hypertension in MHS rats. The lack of significant glomerular hypertrophy and damage in this strain, despite reduced glomerular number, could be related to their (haemodynamic) protection from hypertensive renal disease, possibly due to the hypertrophy of intrarenal arteries. The larger size of glomeruli of MNS rats may be linked to their susceptibility to glomerulosclerosis.

Objective To investigate the value of different biopsy methods for quality evaluation of the donated kidney organ after citizen death. Methods Six cases (6 pairs) of discard donor kidneys were collected from October 2016 to May 2017, respectively, and grouped by wedge biopsy and core needle biopsy. After being fixed and processed for routine paraffin embedding and hematoxylin-eosin (H&E) staining, the specimens were evaluated by the qualified rate of sample, the number of glomerulus, sclerotic glomerulus and small arteries. Results The comparison of two different biopsy methods showed that the number of samples obtained by wedge biopsy was 30. There were 29 samples which were qualified and the qualification rate was 96.7%. The number of samples obtained by core needle biopsy was 30, and only 21 samples were qualified and the qualification rate was 70%. In the wedge biopsy samples, the average number of glomeruli was 22.1 and 6.9 of them were sclerotic glomeruli. The ratio of sclerotic glomeruli was 31.3%. The average number of glomeruli in core needle biopsy samples was 9.5 and 2.1 of them were sclerotic glomeruli. The ratio of sclerotic glomeruli was 22.1%. The average number of arteries in wedge biopsy samples was 5.4, and that in core needle biopsy samples was 3.9. The results indicated that the qualification rate of wedge biopsy was significantly higher than that of core needle biopsy (P<0.01). The number of glomeruli, sclerotic glomeruli and small arteries in wedge biopsy samples was significantly greater than than in core needle biopsy (P<0.05). Conclusion Wedge biopsy was superior to core needle biopsy for the quality evaluation of specimens and identifying clinically significantly histopathological findings. Thus it is potential for wedge biopsy to become the main method in pre-implantation histopathological evaluation. Key words: Kidney transplantation; Histopathological; Renal biopsy

Debrisoquine phenotyping and urinary ochratoxin A in patients with Balkan endemic nephropathy (BEN) and urinary tract tumours (UTT)

Rice is a staple food feeding more than half of the world’s population. Rice disease is one of the major problems affecting rice production. Machine Vision Technology has been used to help develop agricultural production, both in terms of quality and quantity. In this study, convolutional neural network (CNN) was applied to detect and identify diseases in images. We studied 6 varieties of major rice diseases, including blast, bacterial leaf blight, brown spot, narrow brown spot, bacterial leaf streak and rice ragged stunt virus disease. Our studied used well known pre-trained models namely Faster R-CNN, RetinaNet, YOLOv3 and Mask RCNN, and compared their detection performance. The database of rice diseases used in our study contained photographs of rice leaves taken from fields of planting areas. The images were taken under natural uncontrolled environment. We conducted experiments to train and test each model using a total of 6,330 images. The experimental results showed that YOLOv3 provided the best performance in term of mean average precision (mAP) at 79.19% in the detection and classification of rice leaf diseases. The precision obtained from Mask R-CNN, Faster R-CNN, and RetinaNet was at 75.92%, 70.96%, and 36.11%, respectively.

The non-maximum suppression (NMS) algorithm, which merges neighboring bounding boxes as the detection result, is widely utilized in object detection methods. However, the traditional NMS algorithm is not suitable for text detection in natural scenes, especially for long texts and dense texts. In this paper, we observe that the coordinates of candidate bounding boxes are presented in skew distribution. On this observation, an improved NMS algorithm called SD-NMS (Skew Distribution NMS) is designed. First, the mode and the median of the coordinate set are counted to filter out the redundant bounding boxes. Then the left bounding boxes are merged for the location of the text regions. The SD-NMS method improves the detection ability of the model without extra model training and can be easily embedded in text detection methods. The experimental results show that our method obtains F-measure more than other NMS methods in public data sets ICDAR2015 and MSRATD500.

Background and Objectives: Considering the growing shortage of grafts available for kidney transplantation (KT) due to the increase in the number of end-stage renal disease patients, it is essential to utilize all transplantation options. The Karpinski score is a histological scoring system utilized for the evaluation of pre-implantation kidney biopsies from deceased donors. In contrast with Remuzzi Criteria, there has recently been a tendency to perform single KT using grafts with a Karpinski score of 4 or 5. This strategy allows two score 4 or 5 grafts to be used for two different recipients instead of a double-graft KT on one patient. The aim of this study was to analyze the outcomes of single-graft KT with a score of 4 versus 5, and to investigate possible correlations with the clinical characteristics of donors and recipients. Materials and Methods: Retrospective single-centre analysis of 100 KTs performed with a single Karpinski score 4 or 5 graft between January 2014 and December 2022. Results: Grafts with a Karpinski score of 5 harvested from donors older than 70 years of age had a statistically significant (p = 0.014) worse 5-year survival rate (50.0 +/− 18.6%) compared to younger donors (100% for score 5 grafts from donors aged 31–60, and 100% for score 5 grafts from donors aged 61–70). Conversely, donor’s age did not significantly affect the survival of score 4 grafts. Conclusions: The results suggest that a single-graft KT with a Karpinski score 5 graft may be a viable procedure with favourable outcomes. However, for Karpinski score 5 grafts, the use of an older donor beyond the age of 70 seems to be a significant negative factor for the long-term outcome. In such cases, a double KT would potentially be the optimal approach.

Indian traffic sign detection and recognition using deep learning

To identify and locate industrial textile defects accurately, this study proposes a textile detection model based on a convolution neural network (CNN) known as Faster R-CNN. First, a textile defect feature map was extracted by ResNet-101 deep convolution network. Faster R-CNN only extracts features from the last layer of the feature map, which leads to a loss of low-level location information. The proposed method adds the feature pyramid network (FPN) to the network architecture to make an independent prediction for each level in the feature extraction stage. The extracted feature map is input into the regional proposal network, among which the overlapping regional proposals are suppressed. The proposed improved Faster R-CNN model with Region Proposal Network (RPN), Soft Non-Maximum Suppression (NMS), and Region of Interest (ROI) Align can achieve a detection accuracy of 98% and an mean of Average Precision (mAP) of 85%, which is more competitive than the state-of-the-art deep learning-based object detection algorithms.

Detection of traffic signs plays an important role in autonomous driving, traffic surveillance and traffic safety. Previous research in Traffic Sign Detection (TSD) generally focused on traffic signs which are over the roads, the traffic signs on road surface have not been discussed. In this paper, we propose a road surface traffic sign detection system by applying convolutional neural network (CNN). The proposed system consists of two main stages: 1) a hybrid region proposal method to hypothesize the traffic sign locations by taking into account complementary information of color and edge; 2) feature extraction, classification, bounding box regression and non-maximum suppression by Fast R-CNN. Extensive experiments have been conducted using our field-captured dataset, demonstrating outstanding performance with regard to high recall and precision rate. The overall average precision (AP) is about 85.58%.

Background: Aging is characterized by a decline in renal function and by a susceptibility to renal diseases. However it is not clear whether the observed changes are solely hemodynamic, structural or both. We evaluated renal function, functional reserve (RFR) and morphology in healthy elderly individuals. Methods: Healthy participants (n=19) were divided into young (n=6, age range 25-37 years), middle-aged (n=6, 44-74 years) and elderly (n=7, 81-96 years). Nitric oxide (NO), plasma renin activity (PRA) and aldosterone, renal plasma flow (RPF) by p-aminohippurate clearance (CPAH) and glomerular filtration rate (GFR) by inulin clearance (CIN) were determined before and during maximal vasodilating stimuli, induced with the infusion of dopamine and amino acids. Glomerular sclerosis, lumen area and wall thickness of afferent arterioles were determined by kidney biopsy from 36 healthy kidney donors and from 6 nephrectomies for renal carcinoma. Results: GFR and RPF were slightly reduced in elderly individuals whereas filtration fraction (FF) was increased. GFR and RPF did not increase in the elderly after maximal vasodilating stimuli as in young and middle-aged subjects suggesting a reduction of RFR. NO, increased at baseline, did not increase further after vasodilating stimuli; while on the contrary, PRA, similar in the 3 groups at baseline, was not reduced by vasodilating stimuli in the elderly. Sclerotic glomeruli but not glomerular volume were significantly increased by aging. Afferent arteriole lumens were reduced by aging whereas wall thickness was unchanged. Conclusions: Renal function is preserved with aging in healthy subjects at the expense of a complete reduction of RFR. RFR may be wasted to compensate for the increased number of sclerotic glomeruli. Vascular changes, suggested by reduced arteriolar lumen, may be so advanced that even in the presence of high levels of vasodilatory molecules, kidneys are not responsive anymore to maximal vasoactive stimuli.

Clearance experiments were carried out in fetal, young, and adult pigs undergoing salt loading. Kidney sections were histologically examined and the proximal tubular length, glomerular volume, and number of glomeruli were estimated. Between days 48 and 55 of gestation, the significant increase in glomerular filtration rate was proportionately higher than the concentration increase in water, total solutes, and sodium reabsorption. During this period, neither the proximal tubular length nor the glomerular volume of the juxtamedullary nephrons changed significantly. The main morphologic change took place in the superficial cortex, where new glomeruli attached to very short tubules differentiated. From day 55 to birth, the increase in water, total solutes, and sodium filtered loads was proportionately smaller than the concomitant increase in tubular reabsorption. During this phase, the major morphologic change was the significant lengthening of the proximal tubules of the juxtamedullary and superficial nephrons. These findings suggest that for water, sodium, and total solutes the maturation of superficial nephrons is accompanied by a phase of functional glomerular preponderance. Despite the differentiation of the superficial nephrons, renal handling of normally filtered glucose did not change significantly. Functional glomerulotubular coupling for water, sodium, and total solutes took place soon after birth.

Agriculture currently serves as a crucial food source for the global population. However, coconut farming, in particular, demands extensive care and maintenance. This research aims to classify coconut trees across various plantation areas utilizing deep learning techniques, specifically through Faster R-CNN and Mask R-CNN models, based on unmanned aerial vehicle (UAV) imagery. The data collected by both types of RGB UAVs was used for the classification of coconut trees in experimental plots. For the analysis process, aerial photographs obtained from unmanned aerial vehicles, merged with the principles of aerial photography measurement, were analyzed. The research findings revealed that both Faster R-CNN and Mask R-CNN were capable of effectively classifying image data. Nevertheless, to achieve higher accuracy in results, it is essential that the characteristics of the test plots closely align with each other. This study points towards the adoption of a high-resolution tool, ensuring clearer images that facilitate more accurate classification of coconut trees across extensive areas. Consequently, this could lead to more efficient management and maintenance of coconut plantations. Thus, this approach can substantially enhance the efficiency of managing coconut plantations. Doi: 10.28991/HEF-2024-05-04-02 Full Text: PDF

Designing of VehiNet Using Convolutional Neural Networks and Deep Learning Techniques

In view of the time-consuming and laborious manual picking and sorting of strawberries, the direct impact of image recognition accuracy on automatic picking and the rapid development of deep learning(DL), a Faster Regions with Convolutional Neural Network features (R-CNN) strawberry recognition method that combines Mixup data augmentation, a ResNet(Residual Network)50 backbone feature extraction network and a Soft-NMS (Non-Maximum Suppression) algorithm, named the MRS Faster R-CNN, is proposed. In this paper, the transfer learning backbone feature extraction network VGG (Visual Geometry Group) 16 and ResNet50 are compared, and the superior ResNet50 is selected as the backbone network of MRS Faster R-CNN. The data augmentation method of Mixup image fusion is used to improve the learning and generalization ability of the model. The redundant bboxes (bounding boxes) are removed through Soft-NMS to obtain the best region proposal. The freezing phase is added to the training process, effectively reducing the occupation of video memory and shortening the training time. After experimental verification, the optimized model improved the AP (Average Precision) values of mature and immature strawberries by 0.26% and 5.34%, respectively, and the P(Precision) values by 0.81% and 6.34%, respectively, compared to the original model (R Faster R-CNN). Therefore, the MRS Faster R-CNN model proposed in this paper has great potential in the field of strawberry recognition and maturity classification and improves the recognition rate of small fruit and overlapping occluded fruit, thus providing an excellent solution for mechanized picking and sorting.