Attention-adaptive multi-scale feature aggregation dehazing network

In this paper, we propose an end-to-end Attention-adaptive Multi-scale Feature Aggregation Dehazing Network (AMA-Net). The AMA-Net is based on U-Net and designs with three attention-driven modules, Joint Attention Residual Block (JAB), Joint Attention Feature Aggregation Group (JAAG), and Layer Adaptive Attention Feature Aggregation Module (LAA). To be more specific, considering the unevenly distributed haze in images, we introduce the JAB, which adaptively assigns weights to make networks pay attention to important features; to fully utilize the residual features, we propose the residual aggregation (via three JABs) in JAAG; since most feature aggregation methods for dehazing networks do not filter and refine features at different layers, we add LAA to the decoder to weight the features at different layers for aggregation. Through the ablation studies, we verify the effectiveness of the JAB, JAAG, and LAA. Experimental results on synthetic and real-world datasets show that the proposed AMA-Net outperforms relevant state-of-the-art methods.     

Self-Balanced R-CNN for instance segmentation

Current state-of-the-art two-stage models on instance segmentation task suffer from several types of imbalances. In this paper, we address the Intersection over the Union (IoU) distribution imbalance of positive input Regions of Interest (RoIs) during the training of the second stage. Our Self-Balanced R-CNN (SBR-CNN), an evolved version of the Hybrid Task Cascade (HTC) model, brings brand new loop mechanisms of bounding box and mask refinements. With an improved Generic RoI Extraction (GRoIE), we also address the feature-level imbalance at the Feature Pyramid Network (FPN) level, originated by a non-uniform integration between low- and high-level features from the backbone layers. In addition, the redesign of the architecture heads toward a fully convolutional approach with FCC further reduces the number of parameters and obtains more clues to the connection between the task to solve and the layers used. Moreover, our SBR-CNN model shows the same or even better improvements if adopted in conjunction with other state-of-the-art models. In fact, with a lightweight ResNet-50 as backbone, evaluated on COCO minival 2017 dataset, our model reaches 45.3% and 41.5% AP for object detection and instance segmentation, with 12 epochs and without extra tricks. The code is available at https://github.com/IMPLabUniPr/mmdetection/tree/sbr_cnn.     

先验引导的特征金字塔阴影检测网络

阴影检测向来是计算机视觉领域的一个基础性挑战。它需要网络理解图像的全局语义和局部细节信息。本文提出了一种检测阴影区域的先验特征金字塔网络结构。该网络搭建了先验加权模块来提取图像中蕴含的阴影先验信息,通过使用阴影先验信息加权卷积特征,引导网络学习到阴影区域。同时,该网络还应用了特征融合模块来融合粗略的语义信息和自上而下路径中的精细特征,并且加入了后处理,进一步优化网络的预测结果。本文在两个公开的阴影检测基准数据集上进行了实验来评估其网络性能。实验表明,本文的方法能够更准确地检测到阴影,和过去最先进的方法相比也表现出色,在SBU数据集上正确率达到了96.6%,平衡检测错误因子为6.22。      

基于深度学习的轻量化红外弱小车辆目标检测算法研究

伴随高速飞行器的不断发展,目标检测识别作为精确制导的关键一环,需要更高实时性、高准确性地进行目标定位和识别。当前,针对装甲车辆、车辆阵地等时间敏感目标精确检测识别的需求日益迫切,深度学习算法在特征提取及分类器设计上具备优势。文中以特定复杂背景下的小尺寸红外车辆目标为研究对象,针对样本数据少、平台资源受限、实时性要求高、检测精度高等需求,开展基于红外弱小车辆目标检测识别的轻量化深度学习算法研究。项目基于YOLOv5算法进行轻量化剪裁,减小模型的结构,提高实时性;提出了混合域注意力机制模块EPA,该模块通过不降维的局部跨信道交互策略使算法更快速有效地关注重要通道,抑制无效通道,并将通道注意力机制与空间注意力机制结合,使得算法更关注与目标相关的像素信息。提出了残差密集注意模块（RDAB）,该模块由密集残差块与注意力机制EPA构成,通过密集卷积层来提取充分的局部特征,通过注意力机制获取更有效的通道与像素信息,可以使得算法以较小的模型结构获得较好的检测效果。运用设计的网络对数据增广后的小尺寸红外车辆目标数据进行检测识别,并与多种典型算法进行对比实验。由实验结果可知,文中提出的JH-YOLOv5-RDAB网络检测识别效果优于其他网络,权重大小仅为6.6 MB,仅为YOLOv5s算法模型权重的一半,但算法检测效果更优,与93.7 MB的YOLOv5l算法的检测效果接近,mAP50达到95.1%。实验结果表明:该网络在红外弱小车辆目标检测上的优越性和可行性。     

Improved traffic sign recognition algorithm based on YOLOv4-tiny

This study offers an enhanced yolov4-tiny traffic sign identification method for easy deployment on mobile or embedded devices to address the difficulties of a high number of parameters, low recognition accuracy, and poor real-time performance of traffic sign recognition models in complex scenarios. The yolov4-tiny network serves as the model’s foundation. To begin, Octave Convolution is incorporated into the backbone network to eliminate low-frequency feature redundancy, lowering the number of parameters in the model and enhancing computational efficiency. Second, the convolutional block attention module is employed to improve the recognition accuracy of small and medium-sized targets by strengthening the weights of traffic sign regions and suppressing the weights of invalid features. Finally, in the feature fusion stage, the Feature Pyramid Networks structure is replaced with the Simplified Path Aggregation Network structure to improve the fusing of shallow feature information with deep semantic knowledge and lower the miss detection rate even more On the TT100K data set as well as on CCTSDB dataset, the experimental results suggest that our technique can achieve good recognition performance. With a 16MB model size, our solution improves the mean average precision by 3.5 percent and the Frame Per Second by 12.5 f/s when compared to the yolov4-tiny algorithm. Our method outperforms yolov4-tiny in terms of recognition accuracy and detection speed, and it can easily meet the real-time requirements for traffic sign recognition.     

面向电力开关柜的轻量型GB-YOLOv5m状态检测方法

电力开关柜状态灯及仪表具有布局高密、异位同像的特点,从而对边端图像处理技术中的目标形貌、色度对比等基础特征检测能力以及轻量识别能力提出更高要求,为此该文提出一种Ghost-BiFPN-YOLOv5m(GBYOLOv5m)方法。采用加权双向特征金字塔(BiFPN)结构,赋予特征层不同权重以传递更多有效特征信息;增加一个检测层尺度,提升网络对于小目标的检测精度,解决状态灯高密布局引起的小目标识别难问题;利用GhostBottleneck结构替换原主干网络的Bottleneck复杂结构,实现模型的轻量化,为在边端部署模型提供有利条件;通过图像增强技术对有限样本进行状态灯和仪表传递特征的扩充,并通过迁移学习实现算法高速收敛。经10 kV开关柜实测,结果表明该算法对柜体状态灯及仪表共16类目标识别准确率高,均值平均精度(mAP)达97.3%,fps为37.533帧;相较于YOLOv5m算法,在模型大小缩小了37.04%的基础上,mAP提升了10.2%,说明所提方法对灯体与表体的检测能力大幅提升,且轻量识别效率提升明显,对于开关柜电力状态的实时核验与数字孪生信息交互,具有一定的现实意义。     

多尺度特征融合U-Net的遥感影像黑臭水体智能检测

研究采用卫星遥感技术获取高分辨率遥感影像水体样本数据集,基于深度卷积神经网络从高分辨遥感影像中提取水体并进行黑臭水体智能监测,提出了一种改进U-Net的黑臭水体检测网络模型(IWDNet)。基于U-Net结构引入跳跃式多尺度特征融合,结合通道注意力机制、卷积注意力模块、通道与空间注意力机制生成不同多尺度特征融合注意力机制(MFFAM)模块进行对比,并引入空洞卷积扩大网络感受野,最终实现黑臭水体的识别检测。实验证明：基于跳跃式多尺度融合与CBAM注意力机制的黑臭水体检测网络(MFFCBAM-IWNet)模型有效提升了识别精度,在高分辨遥感影像水体样本数据集上表现最佳,总体精度达98.56%,Kappa系数达0.978 4。     

稀疏特征重用的人脸特征提取网络

为了提高人脸特征提取网络的性能,进而提高人脸识别算法的准确率,本文对基于卷积神经网络的人脸特征提取网络进行研究,提出了 SFRNet(Sparse Feature Reuse Network).首先,基于稀疏特征重用、混合特征融合、中心-高斯池化三个创新点,给出了 SFRNet的网络结构.然后,在图像分类数据集Imag...     

YOLSO: You Only Look Small Object

Small object detection is challenging and far from satisfactory. Most general object detectors suffer from two critical issues with small objects: (1) Feature extractor based on classification network cannot express the characteristics of small objects reasonably due to insufficient appearance information of targets and a large amount of background interference around them. (2) The detector requires a much higher location accuracy for small objects than for general objects. This paper proposes an effective and efficient small object detector YOLSO to address the above problems. For feature representation, we analyze the drawbacks in previous backbones and present a Half-Space Shortcut(HSSC) module to build a background-aware backbone. Furthermore, a coarse-to-fine Feature Pyramid Enhancement(FPE) module is introduced for layer-wise aggregation at a granular level to enhance the semantic discriminability. For loss function, we propose an exponential L1 loss to promote the convergence of regression, and a focal IOU loss to focus on prime samples with high classification confidence and high IOU. Both of them significantly improves the location accuracy of small objects. The proposed YOLSO sets state-of-the-art results on two typical small object datasets, MOCOD and VeDAI, at a speed of over 200 FPS. In the meantime, it also outperforms the baseline YOLOv3 by a wide margin on the common COCO dataset.     

Discriminative feature mining hashing for fine-grained image retrieval

With the development of multimedia technology, fine-grained image retrieval has gradually become a new hot topic in computer vision, while its accuracy and speed are limited due to the low discriminative high-dimensional real-valued embedding. To solve this problem, we propose an end-to-end framework named DFMH (Discriminative Feature Mining Hashing), which consists of the DFEM (Discriminative Feature Extracting Module) and SHCM (Semantic Hash Coding Module). Specifically, DFEM explores more discriminative local regions by attention drop and obtains finer local feature expression by attention re-sample. SHCM generates high-quality hash codes by combining the quantization loss and bit balance loss. Validated by extensive experiments and ablation studies, our method consistently outperforms both the state-of-the-art generic retrieval methods as well as fine-grained retrieval methods on three datasets, including CUB Birds, Stanford Dogs and Stanford Cars.     

一种在MR图像中进行脑胶质瘤检测和病灶分割的方法

针对磁共振图像(MRI)进行脑胶质瘤检测及病灶分割对临床治疗方案的选择和手术实施过程的引导都有着重要的价值.为了提高脑胶质瘤的检测效率和分割准确率,该文提出了一种两阶段计算方法.首先,设计了一个轻量级的卷积神经网络,并通过该网络完成MR图像中肿瘤的快速检测及大致定位;接着,通过集成学习过程对肿瘤周围水肿、肿瘤非增强区、...     

基于改进SSD的人群异常行为检测算法研究

针对人群异常行为检测任务中存在的算法复杂度较高,重叠遮挡等带来的检测精度低等问题,本文提出一种基于改进SSD（Single Shot Multi-box Detector）的人群异常行为检测算法。首先采用轻量级网络MobileNet v2代替原始特征提取网络VGG-16,并通过可变形卷积模块构建卷积层来增强感受野,然后通过将位置信息整合到通道注意力中来进行特征增强,能够捕获空间位置之间的远程依赖关系,从而可以较好处理重叠遮挡问题。实验结果表明,本文提出的算法对人群异常行为具有较好的检测效果。     

Joint strong edge and multi-stream adaptive fusion network for non-uniform image deblurring

Non-uniform motion deblurring has been a challenging problem in the field of computer vision. Currently, deep learning-based deblurring methods have made promising achievements. In this paper, we propose a new joint strong edge and multi-stream adaptive fusion network to achieve non-uniform motion deblurring. The edge map and the blurred map are jointly used as network inputs and Edge Extraction Network (EEN) guides the Deblurring Network (DN) for image recovery and to complement the important edge information. The Multi-stream Adaptive Fusion Module (MAFM) adaptively fuses the edge information and features from the encoder and decoder to reduce feature redundancy to avoid image artifacts. Furthermore, the Dense Attention Feature Extraction Module (DAFEM) is designed to focus on the severely blurred regions of blurry images to obtain important recovery information. In addition, an edge loss function is added to measure the difference of edge features between the generated and clear images to further recover the edges of the deblurred images. Experiments show that our method outperforms currently public methods in terms of PSNR, SSIM and VIF, and generates images with less blur and sharper edges.     

基于YOLOv5的轻量级火焰视频流实时检测算法

针对火焰检测模型小目标检测能力差、模型体积大、计算复杂、难以部署到移动端设备的问题,提出了一种轻量化的DGC＿YOLOv5 (you only look once v5)算法。本文首先调用k-means计算函数,计算出适合本文数据集的锚框尺寸;其次引入卷积块注意力机制(convolutional block attention module, CBAM),提高算法对小目标的检测能力;然后利用轻量型的Ghost模块对主干网络中的C3模块进行改进;最后利用深度可分离卷积(depthwise separable convolution, DS＿Conv),用简单的线性计算代替复杂计算,降低模型复杂度,减小模型体积。实验表明,相比原始的YOLOv5算法,本文算法在测试集上的平均精度均值(mean average precision,mAP)可达到94.4%,比原始算法提高1.7个百分点,在视频测试集上平均检测速度可达到71 FPS,可以满足实时检测的要求,参数量和计算量分别减少为原来的41.2%和34.8%,模型大小减少8.4 M,便于后续移动设备端的部署。     

Multi-scale spatial–temporal attention graph convolutional networks for driver fatigue detection

With the development of deep learning, fatigue detection technology for drivers has achieved remarkable achievements. Although the image-based approach achieves good accuracy, it inevitably leads to greater model complexity, which is unsuitable for mobile terminal devices. Luckily, human skeletal data significantly reduces the impact of noise and input data volume while retaining valid information, and it can better deal with real-world driving scenarios with the benefit of robustness in complex driving situations. This paper proposes a lightweight multi-scale spatio-temporal attention graph convolutional network (MS-STAGCN) to efficiently utilize skeleton data to identify driver states by aggregating locally and globally valid face information, which achieves good performance even for lightweight design. The experimental results show that the method achieves 92.4% accuracy on the NTHU-DDD dataset, which can be applied to fatigue detection tasks of the driver in real-world driving scenarios in the future.     

一种基于改进Faster RCNN的通信网光交箱防尘帽智能检测方法

光交箱防尘帽的检测对于通信网络的正常运行具有重要作用。提出了一种基于改进Faster RCNN(Region Convolutional Neural Network)的通信网光交箱防尘帽智能检测方法。首先,对输入图片进行去噪等预处理,通过残差网络(Residual Netwok, ResNet)进行特征提取,并通过区域生成网络(Region Proposal Network, RPN)初步识别出候选区域,然后经过RolAlign进行池化处理,最后经过特征金字塔网络(Feature Pyramid Network, FPN)对光交箱防尘帽进行二次识别。将该方法应用到光交箱防尘帽缺失的智能检测中,取得了很好的效果。     

An Interconnected Feature Pyramid Networks for object detection

Although deep learning makes major breakthroughs in object detection, object detection still faces several limitations listed as follows: (1) Many works underplay the feature selection, leading to the resulting key features are not prominent enough and prone to noise; (2) Many works pass back features in a layer-by-layer manner to achieve multi-scale features. However, as the distance of layers from each other increases, the semantics are diluted, and the transfer of information between layers becomes difficult. To overcome these problems, we propose a new Interconnected Feature Pyramid Networks (IFPN) for feature enhancement. It can simultaneously select attentive features through the attention mechanism and realize the free flow of information. On the basis of the improvements, we design a new IFPN Detector. Experiments on COCO dataset and Smart UVM dataset show that our method can bring a significant improvement.     

基于希尔伯特黄变换和深度卷积神经网络的房颤检测

房颤是一种常见的心律失常,其发病率会随着年龄增长而升高.因此,从心电(ECG)信号中尽早识别出房颤,有助于降低中风风险和心源性死亡率.为有效提高其检测准确率,该文提出一种基于希尔伯特黄变换(HHT)和深度卷积神经网络的房颤检测方法.1维的时域心电信号通过希尔伯特黄变换,转换为时频域信号,旨在通过时频分析,丰富原始信号的...     

Human object interaction detection based on feature optimization and key human-object enhancement

Aiming at the problem of unclear or missing human object interaction behavior objects in complex background, we propose a human object interaction detection algorithm based on feature optimization and key human-object enhancement. In order to solve the problem of missing human behavior objects, we propose Feature Optimized Faster Region Convolutional Neural Network (FOFR-CNN). FOFR-CNN is an object detection network optimized by multi-scale feature optimization algorithm, taking into account both image semantics and image structure. In order to reduce the interference of complex background, we propose a Key Human-Object Enhancement Network. The network uses an instance-based method to enhance the features of interactive objects. In order to enrich the interaction information, we use the graph convolutional network. Experimental results on HICO-DET, V-COCO and HOI-A datasets show that the proposed algorithm has significantly improved accuracy and multi-scale object detection ability compared with other human object interaction algorithms.