Autonomous system refers to a computer system that can operate independently, control itself and make decisions by itself. It does not rely on the support of external systems or software, and can complete various task...
详细信息
Text-generated images are one of the tasks of multimodal machine learning. Although the images generated by previous algorithms can meet the requirements of text description, the generated images are not clear enough ...
详细信息
Considering the challenges associated with robots in optoelectronic imaging applications, typically require real-time and accurate recognition and localization of targets, especially in complex environments. Due to th...
详细信息
Data augmentation has been proven effective for training high-accuracy convolutional neural network classifiers by preventing overfitting. However, building deep neural networks in real-world scenarios requires not on...
详细信息
ISBN:
(纸本)9781713899921
Data augmentation has been proven effective for training high-accuracy convolutional neural network classifiers by preventing overfitting. However, building deep neural networks in real-world scenarios requires not only high accuracy on clean data but also robustness when data distributions shift. While prior methods have proposed that there is a trade-off between accuracy and robustness, we propose IPMix, a simple data augmentation approach to improve robustness without hurting clean accuracy. IPMix integrates three levels of data augmentation (image-level, patch-level, and pixel-level) into a coherent and label-preserving technique to increase the diversity of training data with limited computational overhead. To further improve the robustness, IPMix introduces structural complexity at different levels to generate more diverse images and adopts the random mixing method for multi-scale information fusion. Experiments demonstrate that IPMix outperforms state-of-the-art corruption robustness on CIFAR-C and imageNet-C. In addition, we show that IPMix also significantly improves the other safety measures, including robustness to adversarial perturbations, calibration, prediction consistency, and anomaly detection, achieving state-of-the-art or comparable results on several benchmarks. Code is available at https://***/hzlsaber/IPMix.
Aiming at the problem that convolutional neural networks need to consume a large number of computational units and have a high computational complexity in embedded systems, this paper proposes a Dedicated Hardware Acc...
详细信息
An integrated computerized locking mechanism is vital for upholding the safety of train movements within railway networks. As computer technology progresses swiftly, conventional locking systems are encountering a mul...
详细信息
As one of the common road diseases, the accurate detection of potholes during inspections can help to make timely maintenance measures, which will greatly save road maintenance costs and reduce the incidence of traffi...
详细信息
Virtual reality (VR) conference, as a typical social VR application, has gained popularity in recent years. It offers users located at different locations a fully immersive experience and a sense of togetherness. Howe...
详细信息
Originating from the pistachio tree (Pistacia vera) and prized for their nutritional content and adaptability in the kitchen, pistachios have great financial worth in the agricultural field. Ensuring the quality and s...
详细信息
ISBN:
(纸本)9798331540661;9798331540678
Originating from the pistachio tree (Pistacia vera) and prized for their nutritional content and adaptability in the kitchen, pistachios have great financial worth in the agricultural field. Ensuring the quality and safety of pistachio nuts is absolutely vital and calls for effective industrial post-harvest methods. Modern technology, especially imageprocessing and computer vision algorithms, are augmenting conventional approaches of pistachio categorization and separation. This work intends to build an enhanced classification model for pistachio identification using deep learning, most especially the VGG16 Convolutional Neural Network (CNN). After gathering a dataset of 2,148 high-resolution photos of pistachios, the approach uses resizing and data augmentation methods including rotation, flipping, and zooming to improve the generalization and prevent overfitting of the model. Pretrained on imageNet, the VGG16 model is fine-tuned to fit the particular job of pistachio recognition by means of feature extraction capacity. For feature extraction, the model architecture comprises convolutional layers;for classification, it features fully connected layers. With performance criteria including accuracy, precision, recall, and F1-score watched to guarantee efficient learning, optimization methods including gradient descent and backpropagation are utilized during training. On a test set, the model shows an extraordinary accuracy of 0.98, so highlighting the potential of deep learning models-especially VGG16-in automating and improving quality control systems in agricultural uses. With possible uses in related agricultural research, this highperformance classification model not only meets the demand for effective separation of pistachio species but also considerably increases their economic value.
This paper introduces MicroDeblur, an on-device image motion deblur solution for resource-constrained microcontroller-based vision systems. Although motion blurs caused by the movement or shake of the device (camera) ...
详细信息
ISBN:
(纸本)9798400701184
This paper introduces MicroDeblur, an on-device image motion deblur solution for resource-constrained microcontroller-based vision systems. Although motion blurs caused by the movement or shake of the device (camera) are pervasive in embedded, IoT, and mobile devices, it has been considered a hard nut to crack for many microcontrollers with extremely-limited resources (e.g., hundreds of KB of RAM). To tackle this problem, we combine the DNN (deep neural network) motion deblur method with the classical motion deblur approach and take the best of both worlds, i.e., 1) powerful pattern recognition ability of DNNs and 2) simplicity and stability of matrix-based classical algorithms. To deblur an image, MicroDeblur takes three steps: 1) blur kernel estimation, 2) blur image transformation, and 3) iterative clear image restoration. We propose 1) depth-independent convolution that efficiently estimates the blur kernel (pattern) and 2) Toeplitz-based motion blur modeling that enhances the time and space complexity of the deblurring process by O(n) and O(n(3)), respectively, compared to the existing methods. To the best of our knowledge, MicroDeblur is the first self-sufficient blind deconvolution solution for a stand-alone microcontroller that does not rely on extra hardware or external systems. We implement MicroDeblur on an ARM Cortex-M4F, achieving a competitive quality of deblurred images using 187x and 429x smaller memory and energy, respectively, compared to high-end GPU-based solutions.
暂无评论