The invasion of agricultural diseases and insect pests is a huge difficulty for the growth crops. detection very challenging task. diversity in terms shapes, colors, sizes, as well changes lighting environment, have massive impact on accuracy results. We improved C2F module based DenseBlock proposed DCF to extract low-level features such edge texture diseases. Through sensitivity diseases, can better cope with complex tasks improve robustness detection. background environment different...

To enhance agricultural productivity through the accurate detection of pests under constrained resources mobile devices, we introduce LP-YOLO, a bespoke lightweight object framework optimized for mobile-based insect pest identification. Initially, devise components, namely LP_Unit and LP_DownSample, to serve as direct substitutes majority modules within YOLOv8. Subsequently, develop an innovative attention mechanism, denoted ECSA (Efficient Channel Spatial Attention), which is integrated...

Retinal image processing is very important in the field of clinical medicine.As first step retinal processing, enhancement essential.Because details a are complex and difficult to enhance, we present robust algorithm via dual-tree wavelet transform (DTCWT) morphology-based method this paper.To begin with, utilize pre-processing captured images.Then, DTCWT applied decompose gray obtain high-pass subbands low-pass subbands.Then, Contourlet-based subbands.For subbands, improve morphology...

The development of the agricultural economy is hindered by various pest-related problems. Most pest detection studies only focus on a single category, which not suitable for practical application scenarios. This paper presents deep learning algorithm based YOLOv5, aims to assist workers in efficiently diagnosing information related 102 types pests. To achieve this, we propose new lightweight convolutional module called C3M, inspired MobileNetV3 network. Compared original convolution C3, C3M...

Foggy weather poses significant challenges to outdoor computer vision tasks, such as object detection, by degrading image quality and reducing algorithm reliability. In this paper, we present a novel model for estimating fog density in scenes, aiming enhance detection performance under varying foggy conditions. Using support vector machine (SVM) classification framework, the proposed categorizes unknown images into distinct levels based on both global local fog-relevant features. Key...

Rapid and accurate detection of a C-shaped root canal on mandibular second molars can assist dentists in diagnosis treatment. Oral panoramic radiography is one the most effective methods determining teeth. There are already some traditional based deep learning to learn characteristics tooth images. However, previous studies have shown that accuracy detecting still needs be improved. And it not suitable for implementing these network structures with limited hardware resources. In this paper,...

In recent years, the research and application of ginseng, a famous valuable medicinal herb, has received extensive attention at home abroad. However, with gradual increase in demand for discrepancies are inevitable when using traditional manual method grading appearance quality ginseng. Addressing these challenges was primary focus this study. This study obtained batch ginseng samples enhanced dataset by data augmentation, based on which we refined YOLOv8 network three key dimensions:...

In order to solve the problem of small degree variability between features ginseng grading classes and resulting need for heavy reliance on professionals, this study established a dataset containing 5116 images with three in different contexts proposed ginseng-grading model based an improved ConvNeXt framework. Firstly, Channel Shuffle module was embedded backbone network after down-sampling fully fuse channel improve model’s accuracy. The characterization ability enriched feature space...

With the development of information technology, people access more and rely on network, than 80% in network is replaced by multimedia technology represented images. Therefore, research image processing very important, but most focused a certain aspect. The results unified modeling various aspects are still rare. To this end, paper uses denoising, watermarking, encryption decryption, compression process to carry out modeling, using wavelet transform as method simulate 300 photos from life....

For the analysis of medical images, one most basic methods is to diagnose diseases by examining blood smears through a microscope check morphology, number, and ratio red cells white cells. Therefore, accurate segmentation cell images essential for counting identification. The aim this paper perform smear image combining neural ordinary differential equations (NODEs) with U-Net networks improve accuracy segmentation. In order study effect ODE-solve on speed network, ODE-block module was added...

Monitoring the morphology of blood leukocytes, plays an important role in medical research, especially treatment diseases such as immunodeficiency. Traditional manual detection methods are susceptible to numerous interference factors. Therefore, cells often segmented using deep-learning algorithms. This study proposes a U-Net model based on combination attention mechanism and dilated convolutions. First, traditional convolution double convolutional module network is replaced by convolution,...

Pulmonary nodule recognition is the core module of lung CAD. The Support Vector Machine (SVM) algorithm has been widely used in pulmonary recognition, and Multiple Kernel Learning (MKL-SVM) achieved good results therein. Based on grid search, however, MKL-SVM needs long optimization time course parameter optimization; also its identification accuracy depends fineness grid. In paper, swarm intelligence introduced Particle Swarm Optimization (PSO) combined with to be MKL-SVM-PSO so as realize...

There are still many technological challenges in recognizing small objects complex situations, and the performance of current detection algorithms for is unsatisfactory. Most existing methods mainly use feature pyramid networks to enrich shallow features using contextual features. However, due inconsistency gradients between different layers network, cannot be fully utilized resulting slow improvement object accuracy. To effectively improve algorithm, we propose a new network-based SSRDet....

In the academic world, ginseng (Panax C. A. Meyer) has received much attention as most representative element of Chinese medicine. To address lack traditional algorithms in identification appearance quality and further improve manual on ginseng, we propose a grading method based deep learning, taking advantage benefits learning image identification. Firstly, substituted LeakyReLU for conventional activation function ReLU to enhance predictive power model. Secondly, added an ECA module...

An adaptive optics (AO) system provides real-time compensation for atmospheric turbulence. However, an AO image is usually of poor contrast because the nature imaging process, meaning that contains information coming from both out-of-focus and in-focus planes object, which also brings about a loss in quality. In this paper, we present robust multi-frame restoration algorithm via maximum likelihood estimation. Our proposed uses method with regularization as basic principle, constructs joint...

To accurately detect defects, we propose an enhanced model based on YOLOv8, named STE-YOLO. address the aforementioned challenges, this paper adopts YOLOv8 as improved model. The structure of is follows: We enhance model’s feature extraction and small detail recognition by integrating GhostConv into partial convolutions. In order to attention bias model, introduce a Bottleneck Transformer self-attention convolution layer that effectively improves localization box accuracy. For problem defect...

Text data augmentation is essential in the field of medicine for tasks natural language processing (NLP). However, most traditional text focuses on English datasets, and there little research Chinese datasets to augment sentences. Nevertheless, ignores semantics between words sentences, besides, it has limitations alleviating problem diversity augmented In this paper, a novel medical (MDA) proposed NLP tasks, which combines knowledge graph with generate data. Experiments named entity...

Turbulence generated by random ups and downs in the refractive index of atmosphere produces varying degrees distortion blurring images camera. Traditional methods ignore effect strong turbulence on image. This paper proposes a deep neural network to enhance image clarity under handle this problem. is divided into two sub-networks, generator discriminator, whose functions are mitigate effects determine authenticity recovered After extensive experiments, it proven that present plays role...