This paper presents a novel sub-pixel corner detection algorithm for camera calibration. In order to achieve high accuracy and robust performance, the pixel level candidate regions are firstly identified by Harris detector. Within these regions, the center of gravity (COG) method is used to gain sub-pixel corner detection. Instead of using the intensity value of the regions, we propose to use corner response function (CRF) as the distribution of the weights of COG. The results of camera calibration experiments show that the proposed algorithm is more accurate and robust than traditional COG sub-pixel corner detection methods.
An advanced static VAr generator (ASVG) can be beneficial to power systems in many ways. In order to evaluate its dynamic performance, the device should be modeled appropriately in transient studies. This paper describes the dynamic behavior of ASVG in balanced systems and unbalanced systems. When bus voltage becomes unbalanced due to asymmetrical power systems failure or symmetrical operation, the negative sequence voltage component will cause negative sequence and 3-order harmonic voltage on the AC side of the ASVG. A model is presented for the transient stability studies in unbalanced systems.
This paper focuses on how to predict stock trends quantitatively and to differentiate turning point. Through the summarizing of the main theories and methods in security technical analysis, this paper propose the hypothesis that two contiguous white bars or black bars indicate the uptrend or downtrend, which is called "Two-stage pattern strategy (TSPS)". Empirical test supports the hypothesis and shows that when the forecasted short-term trend is consistent with the bullish or bearish long-term trend, the accurate rate of the hypothesis is higher. Trading rules according to the TSPS method have also been discovered and support the method's profitability.
This paper presents the necessity of device fault diagnosis in a modern industry by taking a condenser as an example and enumerates the fault signs collected from condensers.A microcontroller is designed with a fault diagnosis function by using the backpropagation algorithm.The application is effective.
Scale is a fundamental geographical concept and its role in different geographical contexts has been widely documented. The increasing availability of transport mobility data, in the form of big datasets, enables further incorporation of spatial dependencies and non-stationarity into spatial interaction modeling of transport flows. In this paper a newly developed multiscale flow-focused geographically weighted regression (MFGWR) approach has been applied, in addition to global and local Moran I indices of flow data, to model multiscale socio-economic determinants of regional transport flows between counties across the Jiangsu Province in China. The results have confirmed the power of local Moran I of flow data for identifying urban agglomerations and the effectiveness of MFGWR in exploring multiscale processes of spatial interactions. A comparison between MFGWR and flow-focused geographically weighted regression (FGWR) showed that the MFGWR approach can better interpret the heterogeneous processes of spatial interaction.
Unlike existing works that employ fully-supervised training with polygon annotations, this study proposes an unconstrained text detection system termed Polygon-free (PF), in which most existing polygon-based text detectors (e.g., PSENet [1]) are trained with only upright bounding box annotations. Our core idea is to transfer knowledge from synthetic data to real data to enhance the supervision information of upright bounding boxes. This is made possible with a simple segmentation network, namely Skeleton Attention Segmentation Network (SASN), that includes three vital components (i.e., channel attention, spatial attention and skeleton attention map) and one soft cross-entropy loss.Experiments demonstrate that the proposed Polygon-free yields surprisingly high-quality pixel-level results with only upright bounding box annotations. For example, without using polygon annotations, PSENet achieves an 80.5% F-score on TotalText (vs. 80.9% of fully supervised counterpart), 31.1% better than training directly with upright bounding box annotations, and saves 80%+ labeling costs.
Convolutional neural networks (CNNs) have been proved to be effective models to solve a series of challenging computer vision tasks. However, designing CNN architectures with good performance is still a challenging task. Automatic CNN architecture search algorithms have been proposed in recent years which can find competitive CNN architectures without manual intervention. But automatic search algorithms usually consume considerable computational time and resources. In addition, they only use deep blocks and ignore wide blocks of CNNs, which limits the performance of evolved CNNs. In order to address the above issues, a fast approach for automatically evolving CNN architectures based on deep and wide blocks (FAE-CNN) is proposed through genetic algorithms in this paper. In FAE-CNN, a refined fitness evaluation method based on divided datasets is designed with the purpose to speed up the running time. By introducing the Inception Block, FAE-CNN can evolve optimal CNN architectures from both deep and wide directions. Experimental results on CIFAR10 and CIFAR100 show that FAE-CNN can automatically design CNNs with flexible architectures and better performance in a very short running time.
Robust detection of moving objects in image sequences is an essential part of many vision applications. However, it is not easily achievable with a moving camera since the camera and moving objects motions are mixed together. In this paper we propose a method to detect moving objects under a moving camera. The camera ego-motion is compensated by the corresponding feature sets. The difference image between two consecutive images that ego-motion is compensated is transformed into a binary image using k-means algorithm. According to the clustering results, the region of interest where moving objects are likely to exist is searched by the projection approach. Then local threshold and contour filling methods are applied to detect the accurate moving objects. Experimental results on real image sequences demonstrate that our method can get intact moving objects in the case of a moving camera efficiently.
In this paper, we present a Low-Rank Gradient Estimator (LoGE) to accelerate the finetune-time computation of transformers, especially large language models (LLMs). Unlike Parameter-Efficient Fine-Tuning (PEFT) methods, which primarily aim to minimize the number of fine-tuning parameters, LoGE also significantly reduces the computational load of activation gradient calculations by decomposing pre-trained weights and utilizing low-rank matrices during the backward pass. Our approach includes an effective solution for identifying sensitive and important latent subspaces in large models before training with downstream datasets. As LoGE does not alter the network structure, it can be conveniently integrated into existing models. We validated LoGE’s efficacy through comprehensive experiments across various models on various tasks. For the widely used LLaMA model equipped with LoRA, LoGE achieves up to a 1.3× speedup while maintaining graceful accuracy.
Leaf vein features play an important role in botanical studies. It is also a significant and challenging task in the field of computer vision. This paper proposes a new method of leaf vein extraction and angle measurement. The angle is between the primary vein and the secondary vein. We focus on the leaves whose leaf veins are straight. The algorithm's procedure is as follows: first, through contour extraction we segment foreground and background, get clean images of leaves. Second, change the color space from RGB to HSI, separate the hue component and enhance it. Third, get adaptive thresholds for canny detection based on Otsu algorithm. Then using canny edge detection and Hough transform to detect lines, most of them are leaf veins. We need to select the right lines. And finally we can get the angles between the primary vein and the secondary veins through these lines. Experiments have been conducted with a lot of leaf images. Experimental results show that this method can be used to extract leaf veins and measure angles, even when leaf veins are a little blurry.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)