Explore the vast range of titles available.

A tomato automatic detection method based on an improved YOLOv8s model is proposed to address the low automation level in tomato harvesting in agriculture. The proposed method provides technical support for the automatic harvesting and classification of tomatoes in agricultural production activities. The proposed method has three key components. Firstly, the depthwise separable convolution (DSConv) technique replaces the ordinary convolution, which reduces the computational complexity by generating a large number of feature maps with a small amount of calculation. Secondly, the dual-path attention gate module (DPAG) is designed to improve the model’s detection precision in complex environments by enhancing the network’s ability to distinguish between tomatoes and the background. Thirdly, the feature enhancement module (FEM) is added to highlight the target details, prevent the loss of effective features, and improve detection precision. We built, trained, and tested the tomato dataset, which included 3098 images and 3 classes. The proposed algorithm’s performance was evaluated by comparison with the SSD, faster R-CNN, YOLOv4, YOLOv5, and YOLOv7 algorithms. Precision, recall rate, and mAP (mean average precision) were used for evaluation. The test results show that the improved YOLOv8s network has a lower loss and 93.4% mAP on this dataset. This improvement is a 1.5% increase compared to before the improvement. The precision increased by 2%, and the recall rate increased by 0.8%. Moreover, the proposed algorithm significantly reduced the model size from 22 M to 16 M, while achieving a detection speed of 138.8 FPS, which satisfies the real-time detection requirement. The proposed method strikes a balance between model size and detection precision, enabling it to meet agriculture’s tomato detection requirements. The research model in this paper will provide technical support for a tomato picking robot to ensure the fast and accurate operation of the picking robot.

Existing knowledge distillation (KD) methods are mainly based on features, logic, or attention, where features and logic represent the results of reasoning at different stages of a convolutional neural network, and attention maps symbolize the reasoning process. Because of the continuity of the two in time, transferring only one of them to the student network will lead to unsatisfactory results. We study the knowledge transfer between the teacher-student network to different degrees, revealing the importance of simultaneously transferring knowledge related to the reasoning process and reasoning results to the student network, providing a new perspective for the study of KD. On this basis, we proposed the knowledge distillation method based on attention and feature transfer (AFT-KD). First, we use transformation structures to transform intermediate features into attentional and feature block (AFB) that contain both inference process information and inference outcome information, and force students to learn the knowledge in AFBs. To save computation in the learning process, we use block operations to align the teacher-student network. In addition, in order to balance the attenuation ratio between different losses, we design an adaptive loss function based on the loss optimization rate. Experiments have shown that AFT-KD achieves state-of-the-art performance in multiple benchmark tests.

This paper proposes an improved hybrid algorithm for automated guided vehicles (AGVs) in port environments based on the concept of key obstacles for the JPS and DWA algorithms. Given the complexity of the port environment and the abundance of obstacles, the traditional heuristic function of the JPS algorithm is improved by adding the key obstacle heuristic function. Simultaneously, improvements are made to the evaluation function of the traditional DWA algorithm, where the braking distance is segmented into key obstacle distance and non-key obstacle distance, utilizing the concept of key obstacles. Simulation experiments are conducted using Matlab to demonstrate the effectiveness of the improved algorithm. Moreover, the performance of the hybrid algorithm is compared with five mainstream algorithms in a real simulated port environment, and the final results show the significant enhancement of this paper’s algorithm in several key performance metrics. Thus, this study provides a feasible strategy for improved path planning efficiency for AGV in the port environment.

The large scale and shape variation in breast lesions make their segmentation extremely challenging. A breast ultrasound image segmentation model integrating Mamba-CNN and feature interaction is proposed for breast ultrasound images with a large amount of speckle noise and multiple artifacts. The model first uses the visual state space model (VSS) as an encoder for feature extraction to better capture its long-range dependencies. Second, a hybrid attention enhancement mechanism (HAEM) is designed at the bottleneck between the encoder and the decoder to provide fine-grained control of the feature map in both the channel and spatial dimensions, so that the network captures key features and regions more comprehensively. The decoder uses transposed convolution to upsample the feature map, gradually increasing the resolution and recovering its spatial information. Finally, the cross-fusion module (CFM) is constructed to simultaneously focus on the spatial information of the shallow feature map as well as the deep semantic information, which effectively reduces the interference of noise and artifacts. Experiments are carried out on BUSI and UDIAT datasets, and the Dice similarity coefficient and HD95 indexes reach 76.04% and 20.28 mm, respectively, which show that the algorithm can effectively solve the problems of noise and artifacts in ultrasound image segmentation, and the segmentation performance is improved compared with the existing algorithms.

In response to the challenges faced by traditional LLC resonant converters in simultaneously achieving a wide output voltage and high efficiency, this paper proposes a cascaded DC-DC converter. The front stage of the converter adopts a new LLC topology, and it is cascaded with a boost converter through a reutilization of the output-side inductor. The proposed DC-DC converter leverages the electrical isolation and wide output voltage range advantages of LLC and boost converters, enabling the overall system to achieve broad output voltage regulation and high-efficiency operation. The reutilization of the inductor design further enhances the integration density of the DC-DC converter. A minimum loss collaborative control strategy is introduced for the proposed DC-DC converter. When determining the output voltage and operating the circuit in step-up mode, the duty cycles of the switch in the proposed LLC converter and the switch in the boost converter are adjusted to minimize overall circuit losses while ensuring the rated output voltage. Ultimately, the correctness and practicality of the proposed DC-DC converter and its control strategy are validated through a simulation and an experiment. The overall efficiency of the DC-DC converter can reach up to 94% under optimal conditions.

Previous studies have shown that deep models are often over-parameterized, and this parameter redundancy makes deep compression possible. The redundancy of model weight is often manifested as low rank and sparsity. Ignoring any part of the two or the different distributions of these two characteristics in the model will lead to low accuracy and a low compression rate of deep compression. To make full use of the difference between low-rank and sparsity, a unified framework combining low-rank tensor decomposition and structured pruning is proposed: a hybrid model compression method based on sensitivity grouping (HMC). This framework unifies the existing additive hybrid compression method (AHC) and the non-additive hybrid compression method (NaHC) proposed by us into one model. The latter group the network according to the sensitivity difference of the convolutional layer to different compression methods, which can better integrate the low rank and sparsity of the model compared with the former. Experiments show that our approach achieves a better trade-off between test accuracy and compression ratio when compressing the ResNet family of models than other recent compression methods using a single strategy or additive hybrid compression.

The poor 5-year survival rate for bladder cancers is associated with the lack of efficient diagnostic and treatment techniques. Despite cystoscopy-assisted photomedicine and external radiation being promising modalities to supplement or replace surgery, they remain invasive or fail to provide real-time navigation. Here, we report non-invasive fractionated photodynamic therapy of bladder cancer with full-course real-time near-infrared-II imaging based on engineered X-ray-activated nanotransducers that contain lanthanide-doped nanoscintillators with concurrent emissions in visible and the second near-infrared regions and conjugated photosensitizers. Following intravesical instillation in mice with carcinogen-induced autochthonous bladder tumours, tumour-homing peptide-labelled nanotransducers realize enhanced tumour regression, robust recurrence inhibition, improved survival rates, and restored immune homeostasis under X-ray irradiation with accompanied near-infrared-II imaging. On-demand fractionated photodynamic therapy with customized doses is further achieved based on quantifiable near-infrared-II imaging signal-to-background ratios. Our study presents a promising non-invasive strategy to confront the current bladder cancer dilemma from diagnosis to treatment and prognosis. The poor survival rate for bladder cancers is associated with the lack of effective non-invasive theranostic techniques. Here this group reports a lanthanide-doped nanotransducer activated for real-time NIR-II imaging thereby navigates the photodynamic treatment of bladder cancer.

Immune checkpoint therapy (ICT) provides substantial clinical benefits to cancer patients, but a large proportion of cancers do not respond to ICT. To date, the genomic underpinnings of primary resistance to ICT remain elusive. Here, we performed immunogenomic analysis of data from TCGA and clinical trials of anti-PD-1/PD-L1 therapy, with a particular focus on homozygous deletion of 9p21.3 (9p21 loss), one of the most frequent genomic defects occurring in ~13% of all cancers. We demonstrate that 9p21 loss confers “cold” tumor-immune phenotypes, characterized by reduced abundance of tumor-infiltrating leukocytes (TILs), particularly, T/B/NK cells, altered spatial TILs patterns, diminished immune cell trafficking/activation, decreased rate of PD-L1 positivity, along with activation of immunosuppressive signaling. Notably, patients with 9p21 loss exhibited significantly lower response rates to ICT and worse outcomes, which were corroborated in eight ICT trials of >1,000 patients. Further, 9p21 loss synergizes with PD-L1/TMB for patient stratification. A “response score” was derived by incorporating 9p21 loss, PD-L1 expression and TMB levels in pre-treatment tumors, which outperforms PD-L1, TMB, and their combination in identifying patients with high likelihood of achieving sustained response from otherwise non-responders. Moreover, we describe potential druggable targets in 9p21-loss tumors, which could be exploited to design rational therapeutic interventions. The molecular mechanisms of resistance to immune checkpoint therapy remain elusive. Here, the authors perform immunogenomic analysis of TCGA data and data from clinical trials for antiPD-1/PD-L1 therapy and highlight the association of 9p21 loss with a cold tumor microenvironment and resistance to therapy.

Image dehazing based on convolutional neural networks has achieved significant success; however, there are still some problems, such as incomplete dehazing, color deviation, and loss of detailed information. To address these issues, in this study, we propose a multi-scale dehazing network with dark channel priors (MSDN-DCP). First, we introduce a feature extraction module (FEM), which effectively enhances the ability of feature extraction and correlation through a two-branch residual structure. Second, a feature fusion module (FFM) is devised to combine multi-scale features adaptively at different stages. Finally, we propose a dark channel refinement module (DCRM) that implements the dark channel prior theory to guide the network in learning the features of the hazy region, ultimately refining the feature map that the network extracted. We conduct experiments using the Haze4K dataset, and the achieved results include a peak signal-to-noise ratio of 29.57 dB and a structural similarity of 98.1%. The experimental results show that the MSDN-DCP can achieve superior dehazing compared to other algorithms in terms of objective metrics and visual perception.