Myeloid cyclooxygenase-2/prostaglandin E2/E-type prostanoid receptor 4 promotes transcription factor MafB-dependent inflammatory resolution in acute kidney injury.

Following acute injury to the kidney, macrophages play an important role in recovery of functional and structural integrity, but organ fibrosis and progressive functional decline occur with incomplete recovery. Pro-resolving macrophages are characterized by increased cyclooxygenase 2 (COX-2) expression and this expression was selectively increased in kidney macrophages following injury and myeloid-specific COX-2 deletion inhibited recovery. Deletion of the myeloid prostaglandin E2 (PGE2) receptor, E-type prostanoid receptor 4 (EP4), mimicked effects seen with myeloid COX-2-/- deletion. PGE2-mediated EP4 activation induced expression of the transcription factor MafB in kidney macrophages, which upregulated anti-inflammatory genes and suppressed pro-inflammatory genes. Myeloid Mafb deletion recapitulated the effects seen with either myeloid COX-2 or EP4 deletion following acute kidney injury, with delayed recovery, persistent presence of pro-inflammatory kidney macrophages, and increased kidney fibrosis. Thus, our studies identified a previously unknown mechanism by which prostaglandins modulate macrophage phenotype following acute organ injury and provide new insight into mechanisms underlying detrimental kidney effects of non-steroidal anti-inflammatory drugs that inhibit cyclooxygenase activity.

Deletion of Myeloid Interferon Regulatory Factor 4 (Irf4) in Mouse Model Protects against Kidney Fibrosis after Ischemic Injury by Decreased Macrophage Recruitment and Activation.

BACKGROUND: AKI is characterized by abrupt and reversible kidney dysfunction, and incomplete recovery leads to chronic kidney injury. Previous studies by us and others have indicated that macrophage infiltration and polarization play key roles in recovery from AKI. The role in AKI recovery played by IFN regulatory factor 4 (IRF4), a mediator of polarization of macrophages to the M2 phenotype, is unclear. METHODS: We used mice with myeloid or macrophage cell-specific deletion of Irf4 (MΦ Irf4-/- ) to evaluate Irf4's role in renal macrophage polarization and development of fibrosis after severe AKI. RESULTS: Surprisingly, although macrophage Irf4 deletion had a minimal effect on early renal functional recovery from AKI, it resulted in decreased renal fibrosis 4 weeks after severe AKI, in association with less-activated macrophages. Macrophage Irf4 deletion also protected against renal fibrosis in unilateral ureteral obstruction. Bone marrow-derived monocytes (BMDMs) from MΦ Irf4-/- mice had diminished chemotactic responses to macrophage chemoattractants, with decreased activation of AKT and PI3 kinase and increased PTEN expression. PI3K and AKT inhibitors markedly decreased chemotaxis in wild-type BMDMs, and in a cultured macrophage cell line. There was significant inhibition of homing of labeled Irf4-/- BMDMs to postischemic kidneys. Renal macrophage infiltration in response to AKI was markedly decreased in MΦ Irf4-/- mice or in wild-type mice with inhibition of AKT activity. CONCLUSIONS: Deletion of Irf4 from myeloid cells protected against development of tubulointerstitial fibrosis after severe ischemic renal injury in mice, due primarily to inhibition of AKT-mediated monocyte recruitment to the injured kidney and reduced activation and subsequent polarization into a profibrotic M2 phenotype.

Activation of hypoxia-sensing pathways promotes renal ischemic preconditioning following myocardial infarction.

Ischemic heart disease is the leading cause of death worldwide and is frequently comorbid with chronic kidney disease. Physiological communication is known to occur between the heart and the kidney. Although primary dysfunction in either organ can induce dysfunction in the other, a clinical entity known as cardiorenal syndrome, mechanistic details are lacking. Here, we used a model of experimental myocardial infarction (MI) to test effects of chronic cardiac ischemia on acute and chronic kidney injury. Surprisingly, chronic cardiac damage protected animals from subsequent acute ischemic renal injury, an effect that was accompanied by evidence of chronic kidney hypoxia. The protection observed post-MI was similar to protection observed in a separate group of healthy animals housed in ambient hypoxic conditions prior to kidney injury, suggesting a common mechanism. There was evidence that chronic cardiac injury activates renal hypoxia-sensing pathways. Increased renal abundance of several glycolytic enzymes following MI suggested that a shift toward glycolysis may confer renal ischemic preconditioning. In contrast, effects on chronic renal injury followed a different pattern, with post-MI animals displaying worsened chronic renal injury and fibrosis. These data show that although chronic cardiac injury following MI protected against acute kidney injury via activation of hypoxia-sensing pathways, it worsened chronic kidney injury. The results further our understanding of cardiorenal signaling mechanisms and have implications for the treatment of heart failure patients with associated renal disease.NEW & NOTEWORTHY Experimental myocardial infarction (MI) protects from subsequent ischemic acute kidney injury but worsens chronic kidney injury. Observed protection from ischemic acute kidney injury after MI was accompanied by chronic kidney hypoxia and increased renal abundance of hypoxia-inducible transcripts. These data support the idea that MI confers protection from renal ischemic injury via chronic renal hypoxia and activation of downstream hypoxia-inducible signaling pathways.

Processing neutral tone under the non-attentional condition: a mismatch negativity study.

The neutral tone is a unique tone form in Mandarin as it distinguishes from four canonical tones or full tones on the one hand and integrates phonetic, morphological, syntactical and prosodic information on the other hand. Research to date has been focusing on its unique and variant acoustic features. However, little is known about how native Mandarin speakers process such a unique tone. In the present study, the mismatch negativity was used to explore the comparison-based pre-attentive change detection of Mandarin neutral tone. The mismatch negativity at the time window of 400-800 ms post-first-tone onset was obtained by subtracting event-related potentials to standard neutral tone from event-related potentials to a deviant natural tone. The source analysis of mismatch negativity showed the cortex generator was located at the left temporal lobe. The data suggest that Chinese native speakers process neutral tone automatically under non-attentional conditions, as revealed by the mismatch negativity data aligned with a neutral tone, and that neutral tone does exist as an automatically recognizable one in native Mandarin speakers' tone system.

Spin-lock relaxation rate dispersion reveals spatiotemporal changes associated with tubulointerstitial fibrosis in murine kidney.

PURPOSE: To develop and evaluate a reliable non-invasive means for assessing the severity and progression of fibrosis in kidneys. We used spin-lock MR imaging with different locking fields to detect and characterize progressive renal fibrosis in an hHB-EGFTg/Tg mouse model. METHODS: Male hHB-EGFTg/Tg mice, a well-established model of progressive fibrosis, and age-matched normal wild type (WT) mice, were imaged at 7T at ages 5-7, 11-13, and 30-40 weeks. Spin-lock relaxation rates R1ρ were measured at different locking fields (frequencies) and the resultant dispersion curves were fit to a model of exchanging water pools. The obtained MRI parameters were evaluated as potential indicators of tubulointerstitial fibrosis in kidney. Histological examinations of renal fibrosis were also carried out post-mortem after MRI. RESULTS: Histology detected extensive fibrosis in the hHB-EGFTg/Tg mice, in which collagen deposition and reductions in capillary density were observed in the fibrotic regions of kidneys. R2 and R1ρ values at different spin-lock powers clearly dropped in the fibrotic region as fibrosis progressed. There was less variation in the asymptotic locking field relaxation rate R1ρ∞ between the groups. The exchange parameter Sρ and the inflection frequency ωinfl changed by larger factors. CONCLUSION: Both Sρ and ωinfl depend primarily on the average exchange rate between water and other chemically shifted resonances such as hydroxyls and amides. Spin-lock relaxation rate dispersion, rather than single measurements of relaxation rates, provides more comprehensive and specific information on spatiotemporal changes associated with tubulointerstitial fibrosis in murine kidney.

The Role of the EGF Receptor in Sex Differences in Kidney Injury.

BACKGROUND: Sex differences mediating predisposition to kidney injury are well known, with evidence indicating lower CKD incidence rates and slower decline in renal function in nondiabetic CKD for premenopausal women compared with men. However, signaling pathways involved have not been elucidated to date. The EGF receptor (EGFR) is widely expressed in the kidney in glomeruli and tubules, and persistent and dysregulated EGFR activation mediates progressive renal injury. METHODS: To investigate the sex differences in response to renal injury, we examined EGFR expression in mice, in human kidney tissue, and in cultured cell lines. RESULTS: In wild type mice, renal mRNA and protein EGFR levels were comparable in males and females at postnatal day 7 but were significantly lower in age-matched adult females than in adult males. Similar gender differences in renal EGFR expression were detected in normal adult human kidneys. In Dsk5 mutant mice with a gain-of-function allele that increases basal EGFR kinase activity, males had progressive glomerulopathy, albuminuria, loss of podocytes, and tubulointerstitial fibrosis, but female Dsk5 mice had minimal kidney injury. Oophorectomy had no effect on renal EGFR levels in female Dsk5 mice, while castration protected against the kidney injury in male Dsk5 mice, in association with a reduction in EGFR expression to levels seen in females. Conversely, testosterone increased EGFR expression and renal injury in female Dsk5 mice. Testosterone directly stimulated EGFR expression in cultured kidney cells. CONCLUSIONS: These studies indicate that differential renal EGFR expression plays a role in the sex differences in susceptibility to progressive kidney injury that may be mediated at least in part by testosterone.

Noninvasive quantitative magnetization transfer MRI reveals tubulointerstitial fibrosis in murine kidney.

Excessive tissue scarring, or fibrosis, is a critical contributor to end stage renal disease, but current clinical tests are not sufficient for assessing renal fibrosis. Quantitative magnetization transfer (qMT) MRI provides indirect information about the macromolecular composition of tissues. We evaluated measurements of the pool size ratio (PSR, the ratio of immobilized macromolecular to free water protons) obtained by qMT as a biomarker of tubulointerstitial fibrosis in a well-established murine model with progressive renal disease. MR images were acquired from 16-week-old fibrotic hHB-EGFTg/Tg mice and normal wild-type (WT) mice (N = 12) at 7 T. QMT parameters were derived using a two-pool five-parameter fitting model. A normal range of PSR values in the cortex and outer stripe of outer medulla (CR + OSOM) was determined by averaging across voxels within WT kidneys (mean ± 2SD). Regions in diseased mice whose PSR values exceeded the normal range above a threshold value (tPSR) were identified and measured. The spatial distribution of fibrosis was confirmed using picrosirius red stains. Compared with normal WT mice, scattered clusters of high PSR regions were observed in the OSOM of hHB-EGFTg/Tg mouse kidneys. Moderate increases in mean PSR (mPSR) of CR + OSOM regions were observed across fibrotic kidneys. The abnormally high PSR regions (% area) detected by the tPSR were significantly increased in hHB-EGFTg/Tg mice, and were highly correlated with regions of fibrosis detected by histological fibrosis indices measured from picrosirius red staining. Renal tubulointerstitial fibrosis in OSOM can thus be assessed by qMT MRI using an appropriate analysis of PSR. This technique may be used as an imaging biomarker for chronic kidney diseases.

ErbB4 deletion predisposes to development of metabolic syndrome in mice.

ErbB4, a member of the EGF receptor family, plays a variety of roles in physiological and pathological states. Genetic studies have indicated a link between ErbB4 and type 2 diabetes and obesity, but its role in metabolic syndrome (MetS) has not been reported. In the current study we found that mice with ErbB4 deletion developed MetS after 24 wk on a medium-fat diet (MFD), as indicated by development of obesity, dyslipidemia, hepatic steatosis, hyperglycemia, hyperinsulinemia, and insulin resistance, compared with wild-type mice. ErbB4 deletion mice also exhibited increased amounts of subcutaneous and visceral fat, with increased serum leptin levels, compared with wild-type mice, whereas levels of adiponectin were not significantly different. Histologically, severe inflammation, indicated by F4/80 immunostaining and M1 macrophage polarization, was detected in inguinal and epididymal white adipose tissue in ErbB4 deletion mice. ErbB4 expression decreased during 3T3-L1 preadipocyte differentiation. Administration of neuroregulin 4, a specific ligand for ErbB4, to 3T3-L1 adipocytes had no effect on adipogenesis and lipolysis but significantly inhibited lipogenesis, promoted browning, induced GLUT4 redistribution to the cell membrane, and increased glucose uptake. Neuroregulin 4 also significantly increased glucose uptake in adipocytes isolated from wild-type mice, while these effects were significantly decreased in adipocytes isolated from ErbB4 deletion mice. In conclusion, our results indicate that ErbB4 may play an important role in glucose homeostasis and lipogenesis. ErbB4 deficiency-related obesity and adipose tissue inflammation may contribute to the development of MetS.

Assessment of renal fibrosis in murine diabetic nephropathy using quantitative magnetization transfer MRI.

PURPOSE: Renal fibrosis is a hallmark of progressive renal disease; however, current clinical tests are insufficient for assessing renal fibrosis. Here we evaluated the utility of quantitative magnetization transfer MRI in detecting renal fibrosis in a murine model of progressive diabetic nephropathy (DN). METHODS: The db/db eNOS-/- mice, a well-recognized model of progressive DN, and normal wild-type mice were imaged at 7T. The quantitative magnetization transfer data were collected in coronal plane using a 2D magnetization transfer prepared spoiled gradient echo sequence with a Gaussian-shaped presaturation pulse. Parameters were derived using a two-pool fitting model. A normal range of cortical pool size ratio (PSR) was defined as Mean±2SD of wild-type kidneys (N = 20). The cortical regions whose PSR values exceeded this threshold (threshold PSR) were assessed. The correlations between the PSR-based and histological (collagen IV or picrosirius red stain) fibrosis measurements were evaluated. RESULTS: Compared with wild-type mice, moderate increases in mean PSR values and scattered clusters of high PSR region were observed in cortex of DN mouse kidneys. Abnormally high PSR regions (% area) that were detected by the threshold PSR were significantly increased in renal cortexes of DN mice. These regions progressively increased on aging and highly correlated with histological fibrosis measures, while the mean PSR values correlated much less. CONCLUSION: Renal fibrosis in DN can be assessed by the quantitative magnetization transfer MRI and threshold analysis. This technique may be used as a novel imaging biomarker for DN and other renal diseases.

IL-4/IL-13-mediated polarization of renal macrophages/dendritic cells to an M2a phenotype is essential for recovery from acute kidney injury.

Cytokines IL-4 and IL-13 play important roles in polarization of macrophages/dendritic cells to an M2 phenotype, which is important for recovery from acute kidney injury. Both IL-4 and IL-13 activate JAK3/STAT6 signaling. In mice with diphtheria toxin receptor expression in proximal tubules (selective injury model), a relatively selective JAK3 inhibitor, tofacitinib, led to more severe kidney injury, delayed recovery from acute kidney injury, increased inflammatory M1 phenotype markers and decreased reparative M2 phenotype markers of macrophages/dendritic cells, and development of more severe renal fibrosis after diphtheria toxin administration. Similarly, there was delayed recovery and increased tubulointerstitial fibrosis in these diphtheria toxin-treated mice following tamoxifen-induced deletion of both IL-4 and IL-13, with increased levels of M1 and decreased levels of M2 markers in the macrophages/dendritic cells. Furthermore, deletion of IL-4 and IL-13 led to a decrease of tissue reparative M2a phenotype markers but had no effect on anti-inflammatory M2c phenotype markers. Deletion of IL-4 and IL-13 also inhibited recovery from ischemia-reperfusion injury in association with increased M1 and decreased M2 markers and promoted subsequent tubulointerstitial fibrosis. Thus, IL-4 and IL-13 are required to effectively polarize macrophages/dendritic cells to an M2a phenotype and to promote recovery from acute kidney injury.

Mapping murine diabetic kidney disease using chemical exchange saturation transfer MRI.

PURPOSE: Diabetic nephropathy (DN) is the leading cause of renal failure; however, current clinical tests are insufficient for assessing this disease. DN is associated with changes in renal metabolites, so we evaluated the utility of chemical exchange saturation transfer (CEST) imaging to detect changes characteristic of this disease. METHODS: Sensitivity of CEST imaging at 7 Tesla to DN was evaluated by imaging diabetic mice [db/db, db/db endothelial nitric oxide synthase (eNOS)-/-] that show different levels of nephropathy as well as by longitudinal imaging (8 to 24 weeks). Nondiabetic (db/m) mice were used as controls. RESULTS: Compared with nondiabetic mice, the CEST contrasts of hydroxyl metabolites that correspond to glucose and glycogen were significantly increased in papilla (P), inner medulla (IM), and outer medulla (OM) in db/db and db/db eNOS-/- kidneys at 16 weeks. The db/db eNOS-/- mice that showed advanced nephropathy exhibited greater CEST effects in OM and significant CEST contrasts were also observed in cortex. Longitudinally, db/db mice exhibited progressive increases in hydroxyl signals in IM+P and OM from 12 to 24 weeks and an increase was also observed in cortex at 24 weeks. CONCLUSION: CEST MRI can be used to measure changes of hydroxyl metabolites in kidney during progression of DN. Magn Reson Med 76:1531-1541, 2016. © 2015 International Society for Magnetic Resonance in Medicine.

Alterations of urinary metabolite profile in model diabetic nephropathy.

Countering the diabetes pandemic and consequent complications, such as nephropathy, will require better understanding of disease mechanisms and development of new diagnostic methods. Animal models can be versatile tools in studies of diabetic renal disease when model pathology is relevant to human diabetic nephropathy (DN). Diabetic models using endothelial nitric oxide synthase (eNOS) knock-out mice develop major renal lesions characteristic of human disease. However, it is unknown whether they can also reproduce changes in urinary metabolites found in human DN. We employed Type 1 and Type 2 diabetic mouse models of DN, i.e. STZ-eNOS(-/-) C57BLKS and eNOS(-/-) C57BLKS db/db, with the goal of determining changes in urinary metabolite profile using proton nuclear magnetic resonance (NMR). Six urinary metabolites with significantly lower levels in diabetic compared to control mice have been identified. Specifically, major changes were found in metabolites from tricarboxylic acid (TCA) cycle and aromatic amino acid catabolism including 3-indoxyl sulfate, cis-aconitate, 2-oxoisocaproate, N-phenyl-acetylglycine, 4-hydroxyphenyl acetate, and hippurate. Levels of 4-hydroxyphenyl acetic acid and hippuric acid showed the strongest reverse correlation to albumin-to-creatinine ratio (ACR), which is an indicator of renal damage. Importantly, similar changes in urinary hydroxyphenyl acetate and hippurate were previously reported in human renal disease. We demonstrated that STZ-eNOS(-/-) C57BLKS and eNOS(-/-) C57BLKS db/db mouse models can recapitulate changes in urinary metabolome found in human DN and therefore can be useful new tools in metabolomic studies relevant to human pathology.

Diabetic nephropathy induces alterations in the glomerular and tubule lipid profiles.

Diabetic nephropathy (DN) is a major life-threatening complication of diabetes. Renal lesions affect glomeruli and tubules, but the pathogenesis is not completely understood. Phospholipids and glycolipids are molecules that carry out multiple cell functions in health and disease, and their role in DN pathogenesis is unknown. We employed high spatial resolution MALDI imaging MS to determine lipid changes in kidneys of eNOS(-/-) db/db mice, a robust model of DN. Phospholipid and glycolipid structures, localization patterns, and relative tissue levels were determined in individual renal glomeruli and tubules without disturbing tissue morphology. A significant increase in the levels of specific glomerular and tubular lipid species from four different classes, i.e., gangliosides, sulfoglycosphingolipids, lysophospholipids, and phosphatidylethanolamines, was detected in diabetic kidneys compared with nondiabetic controls. Inhibition of nonenzymatic oxidative and glycoxidative pathways attenuated the increase in lipid levels and ameliorated renal pathology, even though blood glucose levels remained unchanged. Our data demonstrate that the levels of specific phospho- and glycolipids in glomeruli and/or tubules are associated with diabetic renal pathology. We suggest that hyperglycemia-induced DN pathogenic mechanisms require intermediate oxidative steps that involve specific phospholipid and glycolipid species.

A method for small-sized wheat seedlings detection: from annotation mode to model construction.

The number of seedlings is an important indicator that reflects the size of the wheat population during the seedling stage. Researchers increasingly use deep learning to detect and count wheat seedlings from unmanned aerial vehicle (UAV) images. However, due to the small size and diverse postures of wheat seedlings, it can be challenging to estimate their numbers accurately during the seedling stage. In most related works in wheat seedling detection, they label the whole plant, often resulting in a higher proportion of soil background within the annotated bounding boxes. This imbalance between wheat seedlings and soil background in the annotated bounding boxes decreases the detection performance. This study proposes a wheat seedling detection method based on a local annotation instead of a global annotation. Moreover, the detection model is also improved by replacing convolutional and pooling layers with the Space-to-depth Conv module and adding a micro-scale detection layer in the YOLOv5 head network to better extract small-scale features in these small annotation boxes. The optimization of the detection model can reduce the number of error detections caused by leaf occlusion between wheat seedlings and the small size of wheat seedlings. The results show that the proposed method achieves a detection accuracy of 90.1%, outperforming other state-of-the-art detection methods. The proposed method provides a reference for future wheat seedling detection and yield prediction.

Low nitric oxide bioavailability upregulates renal heparin binding EGF-like growth factor expression.

Decreased nitric oxide bioavailability has an important role in the initiation and progression of diabetic nephropathy, but the underlying mechanisms remain unclear. Here, we found that heparin-binding epidermal growth factor-like growth factor (HB-EGF) expression levels increased in the kidneys of both endothelial nitric oxide synthase (eNOS)-knockout and eNOS-knockout diabetic (Lepr(db/db)) mice as early as at 8 weeks of age. Further increases in expression were only seen in eNOS-knockout diabetic mice and paralleled the progression of glomerulopathy. HB-EGF expression increased in endothelium, podocytes, and tubular epithelial cells. In cultured glomerular endothelial cells, the nitric oxide synthase inhibitors NG-nitro-L-arginine methyl ester (L-NAME) or L-N5-(1-iminoethyl) ornithine increased HB-EGF protein expression. Administration of L-NAME dramatically increased renal HB-EGF expression and urinary HB-EGF excretion in diabetic mice. On the other hand, replenishing nitric oxide with sodium nitrate in eNOS-knockout diabetic mice reduced urinary HB-EGF excretion and inhibited the progression of diabetic nephropathy. Furthermore, specific deletion of HB-EGF expression in the endothelium attenuated renal injury in diabetic eNOS-knockout mice. Thus, our results suggest that decreased nitric oxide bioavailability leads to increased HB-EGF expression, which may be an important mediator of the resulting progressive diabetic nephropathy in eNOS-knockout diabetic mice.

Control of charge transport in electronically active systems towards integrated biomolecular circuits (IbC).

The miniaturization of traditional silicon-based electronics will soon reach its limitation as quantum tunneling and heat become serious problems at the several-nanometer scale. Crafting integrated circuits via self-assembly of electronically active molecules using a "bottom-up" paradigm provides a potential solution to these technological challenges. In particular, integrated biomolecular circuits (IbC) offer promising advantages to achieve this goal, as nature offers countless examples of functionalities entailed by self-assembly and examples of controlling charge transport at the molecular level within the self-assembled structures. To this end, the review summarizes the progress in understanding how charge transport is regulated in biosystems and the key redox-active amino acids that enable the charge transport. In addition, charge transport mechanisms at different length scales are also reviewed, offering key insights for controlling charge transport in IbC in the future.

Improving multi-scale detection layers in the deep learning network for wheat spike detection based on interpretive analysis.

BACKGROUND: Detecting and counting wheat spikes is essential for predicting and measuring wheat yield. However, current wheat spike detection researches often directly apply the new network structure. There are few studies that can combine the prior knowledge of wheat spike size characteristics to design a suitable wheat spike detection model. It remains unclear whether the complex detection layers of the network play their intended role. RESULTS: This study proposes an interpretive analysis method for quantitatively evaluating the role of three-scale detection layers in a deep learning-based wheat spike detection model. The attention scores in each detection layer of the YOLOv5 network are calculated using the Gradient-weighted Class Activation Mapping (Grad-CAM) algorithm, which compares the prior labeled wheat spike bounding boxes with the attention areas of the network. By refining the multi-scale detection layers using the attention scores, a better wheat spike detection network is obtained. The experiments on the Global Wheat Head Detection (GWHD) dataset show that the large-scale detection layer performs poorly, while the medium-scale detection layer performs best among the three-scale detection layers. Consequently, the large-scale detection layer is removed, a micro-scale detection layer is added, and the feature extraction ability in the medium-scale detection layer is enhanced. The refined model increases the detection accuracy and reduces the network complexity by decreasing the network parameters. CONCLUSION: The proposed interpretive analysis method to evaluate the contribution of different detection layers in the wheat spike detection network and provide a correct network improvement scheme. The findings of this study will offer a useful reference for future applications of deep network refinement in this field.

Small and Oriented Wheat Spike Detection at the Filling and Maturity Stages Based on WheatNet.

Accurate wheat spike detection is crucial in wheat field phenotyping for precision farming. Advances in artificial intelligence have enabled deep learning models to improve the accuracy of detecting wheat spikes. However, wheat growth is a dynamic process characterized by important changes in the color feature of wheat spikes and the background. Existing models for wheat spike detection are typically designed for a specific growth stage. Their adaptability to other growth stages or field scenes is limited. Such models cannot detect wheat spikes accurately caused by the difference in color, size, and morphological features between growth stages. This paper proposes WheatNet to detect small and oriented wheat spikes from the filling to the maturity stage. WheatNet constructs a Transform Network to reduce the effect of differences in the color features of spikes at the filling and maturity stages on detection accuracy. Moreover, a Detection Network is designed to improve wheat spike detection capability. A Circle Smooth Label is proposed to classify wheat spike angles in drone imagery. A new micro-scale detection layer is added to the network to extract the features of small spikes. Localization loss is improved by Complete Intersection over Union to reduce the impact of the background. The results show that WheatNet can achieve greater accuracy than classical detection methods. The detection accuracy with average precision of spike detection at the filling stage is 90.1%, while it is 88.6% at the maturity stage. It suggests that WheatNet is a promising tool for detection of wheat spikes.