image decomposition, which separates a given input image into structure and texture images, has been used for various applications in the fields of computer graphics and imageprocessing. Most previous image decomposi...
详细信息
image decomposition, which separates a given input image into structure and texture images, has been used for various applications in the fields of computer graphics and imageprocessing. Most previous image decomposition algorithms and applications start with high-quality images, but it requires numerous steps to separate and edit the high-resolution structure and texture images from the low-resolution input counterparts. This paper proposes a simple but effective end-to-end deep neural image decomposition network, which is called "scalable image decomposition", by decomposing and upscaling structure and texture images from the degraded input image at the same time. We train the deep neural network to automatically estimate high-resolution structure and texture from the low-resolution input image by designing a shared feature extractor, and structure and texture upscaling networks which have a powerful capability to establish and distinguish a complex mapping between the low-resolution input image and high-quality structure and texture, while preserving more contextual information without any prior information regarding the low-resolution input image. Quantitative and qualitative analyses of the proposed scalable image decomposition network validate that the proposed method is stable and robust against blurring and staircase effects by separating texture and structure upscaling networks in real-time. The predicted upscaled structure and texture images can be used to a variety of applications, such as image abstraction, detail enhancement, and pencil sketching.
In the past decade, various haze removal techniques have been widely reported for object recognition. But hitherto little has been identified on the use of single image dehazing using transfer learning approach for ob...
In the past decade, various haze removal techniques have been widely reported for object recognition. But hitherto little has been identified on the use of single image dehazing using transfer learning approach for object detection. Single image dehazing is an emerging computer vision technology which offers some of the extreme benefits over the existing techniques such as consumes less processingtime, requires less space for realtime dehaze purpose, etc. In this study, we combine both the object detection and image dehazing methods for realtime applications such as-remote sensing, video surveillances, driverless automatic vehicles, etc. This paper presents an effective and efficient image dehazing method using transfer learning which helps to recognize objects in realtime with more clarity and that can automatically detect objects with a high recognition rate and lesser probability of error. Our tests show that object detection becomes less accurate as the haze intensity increases; yet, under all haze circumstances (low, medium, or heavy), our jointly trained model AOD-net +YOLO v3 consistently outperforms non-joint and naïve YOLO v3 techniques.
Pneumothorax is a common pulmonary disease that can lead to dyspnea and can be life-threatening. X-ray examination is the main means to diagnose this disease. Computer-aided diagnosis of pneumothorax on chest X-ray, a...
详细信息
Pneumothorax is a common pulmonary disease that can lead to dyspnea and can be life-threatening. X-ray examination is the main means to diagnose this disease. Computer-aided diagnosis of pneumothorax on chest X-ray, as a prerequisite for a timely cure, has been widely studied, but it is still not satisfactory to achieve highly accurate results. In this paper, an image classification algorithm based on the deep convolutional neural network (DCNN) is proposed for high-resolution medical image analysis of pneumothorax X-rays, which features a Network In Network (NIN) for cleaning the data, random histogram equalization data augmentation processing, and a DCNN. The experimental results indicate that the proposed method can effectively increase the correct diagnosis rate of pneumothorax, and the Area under Curve (AUC) of the test verified in the experiment is 0.9844 on ZJU-2 test data and 0.9906 on the ChestX-ray14, respectively. In addition, a large number of atmospheric pleura samples are visualized and analyzed based on the experimental results and in-depth learning characteristics of the algorithm. The analysis results verify the validity of feature extraction for the network. Combined with the results of these two aspects, the proposed X-ray imageprocessing algorithm can effectively improve the classification accuracy of pneumothorax photographs.
Recently, the energy-efficient photometric stereo method using an event camera (EventPS [67]) has been proposed to recover surface normals from events triggered by changes in logarithmic Lambertian reflections under a...
详细信息
Non-rigid registration of cell nuclei in time-lapse microscopy images needs accurate estimation of the deformation fields between the reference and all other images. To address the issue of accumulated errors in class...
详细信息
ISBN:
(数字)9798331520526
ISBN:
(纸本)9798331520533
Non-rigid registration of cell nuclei in time-lapse microscopy images needs accurate estimation of the deformation fields between the reference and all other images. To address the issue of accumulated errors in classical temporal incremental registration approaches, we introduce a new semi-incremental optimization method. Based on a proper initialization constructed through motion concatenation, which exploits temporal coherence within the image sequence, the deformation field between each image and the reference is computed, reducing accumulated errors and yielding more reliable results. Experiments on realtime-lapse cell images demonstrate that our method outperforms previous approaches, including deeplearning models, in terms of registration accuracy. Additionally, computation on a GPU significantly enhances efficiency.
Roads are transport infrastructure, and regular maintenance ensures safety and efficiency. Deteriorating asphalt can lead to potholes, increasing the risk of accidents. Therefore, this paper evaluates the deep learnin...
详细信息
ISBN:
(数字)9798331518097
ISBN:
(纸本)9798331518103
Roads are transport infrastructure, and regular maintenance ensures safety and efficiency. Deteriorating asphalt can lead to potholes, increasing the risk of accidents. Therefore, this paper evaluates the deeplearning architecture for detecting potholes and facilitating lane change assistance, involving federated learning and edge computing for road safety improvement. The pothole detection module uses CNNs trained on various road image datasets to identify potholes. Federated learning allows vehicles to train the model locally using their data and share updated model parameters with a central server, preserving privacy and reducing data transmission. Integration of TensorFlow and Keras ensures efficient model deployment and high accuracy in pothole detection. Once potholes are detected, it is activated by the lane change assistance system that aids accident prevention. Edge computing locally processes data, ensuring that the decisions and lane changes occur without delays in realtime. The models are continuously improved because data from several vehicles contribute to a more robust system with federated learning. This approach improves pothole detection but also enhances lane change assistance, reducing accident risks. Federated learning is ensured to be privacy-preserving, and edge computing ensures fast processing. The experimental results demonstrate that the system surpasses the existing models in accuracy, safety, and real-time performance.
Defects or cracks in roads, building walls, floors, and product surfaces can degrade the completeness of the product and become an impediment to quality control. Machine learning can be a solution for detecting defect...
详细信息
Defects or cracks in roads, building walls, floors, and product surfaces can degrade the completeness of the product and become an impediment to quality control. Machine learning can be a solution for detecting defects effectively without human experts;however, the low-power computing device cannot afford that. In this paper, we suggest a crack detection system accelerated by edge computing. Our system consists of two: Rsef and Rsef-Edge. Rsef is a real-time segmentation method based on effective feature extraction that can perform crack image segmentation by optimizing conventional deeplearning models. Then, we construct the edge-based system, named Rsef-Edge, to significantly decrease the inference time of Rsef, even in low-power IoT devices. As a result, we show both a fast inference time and good accuracy even in a low-powered computing environment.
image contrast in multispectral optoacoustic tomography can be reduced by electrical noise. We present a deeplearning method to remove electrical noise from optoacoustic signals and thereby significantly enhance morp...
详细信息
ISBN:
(数字)9781510647138
ISBN:
(纸本)9781510647138;9781510647121
image contrast in multispectral optoacoustic tomography can be reduced by electrical noise. We present a deeplearning method to remove electrical noise from optoacoustic signals and thereby significantly enhance morphological and spectral contrast.
To address the issues of low real-time performance and poor algorithm accuracy in detecting miner behavior underground, we propose a high-precision real-time detection method named DSY-YOLOv8n based on the characteris...
详细信息
To address the issues of low real-time performance and poor algorithm accuracy in detecting miner behavior underground, we propose a high-precision real-time detection method named DSY-YOLOv8n based on the characteristics of human body behavior. This method integrates DSConv into the backbone network to enhance multi-scale feature extraction. Additionally, SCConv-C2f replaces C2f modules, reducing redundant calculations and improving model training speed. The optimization strategy of the loss function is employed, and MPDIoU is used to improve the model's accuracy and speed. The experimental results show: (1) With almost no increase in parameters and calculation amount, the mAP50 of the DSY-YOLOv8n model is 97.4%, which is a 3.2% great improvement over the YOLOv8n model. (2) Compared to Faster-R-CNN, YOLOv5s, and YOLOv7, DYS-YOLOv8n has improved the average accuracy to varying degrees while significantly increasing the detection speed. (3) DYS-YOLOv8n meets the real-time requirements for behavioral detection in mines with a detection speed of 243FPS. In summary, the DYS-YOLOv8n offers a real-time, efficient, and lightweight method for detecting miner behavior in mines, which has high practical value.
暂无评论