ABSTRACT
In order to address the challenges of inefficiency and insufficient accuracy in the manual identification of young citrus fruits during thinning processes, this study proposes a detection methodology using the you only look once for complex backgrounds of young citrus fruits (YCCB-YOLO) approach. The method first constructs a dataset containing images of young citrus fruits in a real orchard environment. To improve the detection accuracy while maintaining the computational efficiency, the study reconstructs the detection head and backbone network using pointwise convolution (PWonv) lightweight network, which reduces the complexity of the model without affecting the performance. In addition, the ability of the model to accurately detect young citrus fruits in complex backgrounds is enhanced by integrating the fusion attention mechanism. Meanwhile, the simplified spatial pyramid pooling fast-large kernel separated attention (SimSPPF-LSKA) feature pyramid was introduced to further enhance the multi-feature extraction capability of the model. Finally, the Adam optimization function was used to strengthen the nonlinear representation and feature extraction ability of the model. The experimental results show that the model achieves 91.79% precision (P), 92.75% recall (R), and 97.32% mean average precision (mAP)on the test set, which were improved by 1.33%, 2.24%, and 1.73%, respectively, compared with the original model, and the size of the model is only 5.4 MB. This study could meet the performance requirements for citrus fruit identification, which provides technical support for fruit thinning.
ABSTRACT
Apple leaf diseases without timely control will affect fruit quality and yield, intelligent detection of apple leaf diseases was especially important. So this paper mainly focuses on apple leaf disease detection problem, proposes a machine vision algorithm model for fast apple leaf disease detection called LALNet (High-speed apple leaf network). First, an efficient sacked module for apple leaf detection, known as EALD (efficient apple leaf detection stacking module), was designed by utilizing the multi-branch structure and depth-separable modules. In the backbone network of LALNet, (High-speed apple leaf network) four layers of EALD modules were superimposed and an SE(Squeeze-and-Excitation) module was added in the last layer of the model to improve the attention of the model to important features. A structural reparameterization technique was used to combine the outputs of two layers of deeply separable convolutions in branch during the inference phase to improve the model's operational speed. The results show that in the test set, the detection accuracy of the model was 96.07%. The total precision was 95.79%, the total recall was 96.05%, the total F1 was 96.06%, the model size was 6.61 MB, and the detection speed of a single image was 6.68 ms. Therefore, the model ensures both high detection accuracy and fast execution speed, making it suitable for deployment on embedded devices. It supports precision spraying for the prevention and control of apple leaf disease.
ABSTRACT
Objective:To explore the accuracy and repeatability of the smartphone scoliosis screening APP developed by our team in the measurement of the Cobb angle of adolescent idiopathic scoliosis (AIS).Methods:Clinical data of 60 patients with AIS admitted to Department of Spine and Spinal Surgery, Henan Province People′s Hospital from August 2020 to February 2021 were analyzed retrospectively.Three surveyors measured the coronal main curvature Cobb angle, sagittal thoracic kyphosis (TK) angle, thoracolumbar kyphosis (TLK) angle and lumbar kyphosis (LL) angle on whole-spine lateral X-ray films of 60 patients with AIS by means of the protractor, scoliosis screening APP and computerized Picture Archiving and Communication System (PACS). The time and results of each measurement were recorded.The measurement was repeated once after 2 weeks.Paired t-test was used to compare the measurement time of the APP method and the protractor method.Taking the Cobb angle measurement results of the PACS system as the reference standard, the accuracy of Cobb angle measurement by the APP method was analyzed by paired t-test.The repeatability of the surveyor and the consistency between the surveyors was compared by intraclass correlation coefficient (ICC). Results:Among the 60 patients with AIS, there were 17 males and 43 females, aged from 10 to 16 years [(12.2±2.4) years]. The main bends were thoracic curvature (Lenke Ⅰ) in 23 cases, Lenke Ⅱ in 18 cases and thoracolumbar curvature/lumbar curvature (Lenke V) in 19 cases.The APP method took significantly less time to measure the Cobb angle than the protractor method ( P<0.05). There was no significant difference in the Cobb angle measured by the APP method and PACS method ( P>0.05). The results of the coronal main curvature Cobb angle, TK angle, TLK angle and LL angle measured by 3 surveyors through the APP method were all in good agreement (ICC=0.990, 0.988, 0.986, 0.987). The repeatability (ICC 0.973-0.982) of the coronal main curvature Cobb angle, TK, TLK and LL measured twice before and after the APP method were both better that of the protractor method (ICC 0.933-0.954). Conclusions:Compared with the traditional protractor, the smartphone scoliosis screening APP has the advantages of short measurement time, high efficiency, excellent accuracy and good repeatability in measuring the Cobb angle of AIS.
ABSTRACT
Objective:To explore the clinical effect of the 3D printing pre-installed screw channel model in assisting screw placement of single complete segmented congenital hemivertebrectomy.Methods:Clinical data of 13 children treated with single complete segmented congenital hemivertebrectomy in the Department of Spine and Spinal Surgery of Henan Provincial People′s Hospital from August 2016 to January 2019 were retrospectively analyzed.Among them, there were 5 males and 8 females with the mean age of 9.9 (5-14) years.Categorized by the lesion location, 3 cases were located at T 9, 2 cases at T 10, 5 cases at T 11, 1 case at T 12, and 2 cases at L 1.During the operation, the 3D printing pre-installed screw channel model was used to assist the placement of pedicle screws.The accuracy of screw placement was assessed by the postoperative CT.All children were routinely examined by full-length anterior and lateral X-ray of spine in the standing position before and after surgery to measure the Cobb angles at the coronal and sagittal view.Furthermore, the correction rate of scoliosis and kyphosis after surgery and during follow-up was also calculated.The One-Way repeated measures ANOVA was used to compare the Cobb angle of scoliosis and kyphosis before surgery, after surgery and during follow-up. Results:A total of 85 pedicle screws were placed in 13 children, with the accuracy rate of screw placement of 95.3%.The mean surgery time and intraoperative blood loss were (216.9±28.3) min, and (478.5±132.6) mL, respectively.Scoliosis Cobb was corrected from (57.1±12.7)° to (12.7±4.7)° with a correction rate of (78.4±5.9)%, which was (14.2±7.0)° at the last follow-up.Kyphosis angle was corrected from (46.2±8.4)° to (13.2±4.4)° with a correction rate of (72.6±7.0)%, which was (14.0±3.4)° at the last follow-up.None of the children had serious complications like vascular and nerve damage.The mean postoperative follow-up was 12.3 (6-18) months.No significant loss of angle was detected during the follow-up period.There were significant differences in the lateral and kyphotic angles after surgery and during follow-up compared with preoperative ones (all P<0.05). No significant difference was detected between the postoperative lateral and kyphotic angles and those at the last follow-up (all P>0.05). Conclusion:The 3D printing pre-installed screw channel model used to assist screw placement of single complete segmented congenital hemivertebrectomy can improve the precision of screw placement and the orthopedic effect on lateral kyphosis.
