Mean Shift Method Research Articles

MNN: Mixed nearest-neighbors for self-supervised learning

In contrastive self-supervised learning, positive samples are typically drawn from the same image but in different augmented views, resulting in a limited source of positive samples. An effective way to alleviate this problem is to incorporate the relationship between samples, which involves including the top-K nearest neighbors of positive samples. However, the issue of false neighbors (i.e., neighbors that do not belong to the same category as the positive sample) is a significant yet often overlooked challenge. In this paper, we present a simple self-supervised learning framework called Mixed Nearest-Neighbors (MNN) for Self-Supervised Learning. The primary objective of the MNN is to enhance the robustness and accuracy of the model by strategically incorporating synthesized neighboring samples. Specifically, our proposed method utilizes image mixture to mitigate the effects of noise introduced by false neighbors, while adopting an intuitive weighting strategy that effectively integrates these synthesized neighbors into the model. The linear evaluation of MNN on CIFAR-10, CIFAR,100, STL-10, and Tiny ImageNet shows an improvement of 0.41% to 1.96% over the previous state-of-the-art (SOTA) methods and 1.82% to 8.8% over the Mean Shift (MSF) method. Furthermore, the proposed components can be seamlessly integrated into other self-supervised learning algorithms to enhance their performance. Code is available at https://github.com/pc-cp/MNN.

Automated inspection approach for OCA particles in multi-layered cover glass of display module assembly

Abstract CG (Cover Glass) is a critical component of electronic screens, and its manufacturing quality is closely relevant to the display effect. To satisfy the requirement of quality control, a machine vision system for real-time inspection of particles in the OCA (Optically Clear Adhesive) layer of CG is presented in this paper. With a brief description of the optical logic of particle imaging and the design of the vision system, our emphasis is put on the post-processing image analysis. To align particles regions in the CG under multi-mode imaging, a spatial alignment calibration algorithm is proposed with perspective distortion correction to calculate the triaxial offset. Then, a CLAHE+PM filtering is adopted to enhance the contrast of the particle. Furthermore, a Meanshift method combined with adaptive local thresholding is proposed to extract the contours of tiny particles. Finally, to distinguish between multiple layers of particles in the CG and detect OCA particles, a combination of fast background reconstruction and Averaged Stochastic Gradient Descent-Support Vector Machine is used. According to in-line experiments and tests, our system can find out a majority of the OCA particles with a PR (over-detection rate) of 1.31% and a PM (miss detection rate) of 0.33% for over 10 000 CG samples.

Phase unwrapping error identification and suppression method in φ-OTDR systems based on PELT-VMD-ARIMA.

In phase-sensitive optical time-domain reflectometer (φ-OTDR) systems, phase unwrapping errors can distort vibration information. To address this issue, a phase unwrapping error identification and suppression method combining pruned exact linear time (PELT) changepoint detection, variational mode decomposition (VMD), and autoregressive integrated moving average (ARIMA) models, termed PELT-VMD-ARIMA, is proposed. Firstly, the principle of the proposed method is introduced, and its effectiveness is verified through a series of numerical simulation experiments. Next, piezoelectric transducers (PZTs) are employed as seismic sources in experiments involving single-frequency and chirp signals. Compared to the mean-shift method, the proposed method reduces the average root mean square error (RMSE) by 70.36% within 2δ range around the changepoints. Finally, the proposed method was validated through an active source seismic application. The results demonstrate the efficacy of the proposed method in identifying and suppressing phase unwrapping errors, thereby enhancing signal quality. This method enhances the vibration recognition capability of φ-OTDR systems, which facilitates precise distributed acoustic sensing applications.

Comparison of the Direct Supervised Classification and Segment Mean Shift Classification: A Case Study of Hetauda Sub-Metropolitan City of Nepal

Comparison between Land Use Land Cover (LULC) classification and methods is a common practice to assess their suitability and reliability in specific areas. In the case of Hetauda Sub-Metropolitan City in Nepal, Sentinel-2B satellite imagery was utilized to evaluate the accuracy of LULC classification through both direct supervised classification and segment mean shift classification. Direct supervised classification was conducted in ERDAS Imagine, while segment mean shift classification was performed in ArcGIS to identify six LULC classes: forest, agriculture, built-up area, bare area, grassland, and waterbody. Pre-processing steps such as layer stacking, mosaicking, data extraction, atmospheric and radiometric correction were employed to enhance the imagery. Training samples were chosen via visual interpretation based on local knowledge to train the Maximum Likelihood Classifier (MLC). The results revealed that the segment mean shift method achieved the highest overall accuracy at 90%, whereas direct supervised classification yielded 85%. Furthermore, kappa statistics indicated a strong level of agreement, with segment mean shift scoring the highest at 0.88 and direct supervised classification at 0.81. Individual class accuracies varied, with forest classification being the most reliable at 98% and waterbody classification at 71%. Overall, the segment mean shift method was deemed more accurate and suitable than the pixel-based direct supervised classification for LULC classification in Hetauda Sub-Metropolitan City, Nepal.

A framework to identify guano on photovoltaic modules in offshore floating photovoltaic power plants

Guano are an important factor affecting the cleanliness of photovoltaic modules on floating solar power plants at sea. It can lead to a decrease in photoelectric conversion efficiency, power loss, and even the occurrence of “hot spots”, thereby causing damage to the components. Therefore, the segmentation and detection of guano are crucial for visual automation in the cleaning and inspection processes. However, the composition, density, and thickness of guano naturally vary, leading to inconsistent levels of transparency and color. The uneven intensity of guano images greatly reduces the accuracy of segmentation and detection. Addressing this issue, this study proposes a segmentation algorithm based on combining different color channels to segment guano on the surface of photovoltaic modules. The mean shift method is used for adaptive segmentation to facilitate the detection of guano. Furthermore, the segmentation results obtained by this method are introduced into the input space to improve the traditional Mask RCNN. In addition, this study successfully produced a dataset of guano on the surface of photovoltaic modules using on-site data collection and with the help of a platform built in-house in the laboratory. The experimental results on the self-constructed dataset demonstrate that the enhanced Mask R-CNN model has shown an approximate increase of 5.9% and 6.0% in mAP values for object recognition and segmentation compared to the traditional Mask R-CNN model. This indicates the effectiveness of the methodology proposed in this study.

Visual reconstruction method of architectural space under laser point cloud big data

In order to solve the problem that the reconstruction accuracy and integrity are affected due to the large amount of point cloud data in the process of building space reconstruction, the visual reconstruction method of building space under laser point cloud big data is studied. The three-dimensional laser scanner is used to collect the laser point cloud big data in the building space, and the laser point cloud big data is organized and processed through three steps: hierarchical calculation of the point cloud pyramid, thinning treatment and block treatment. From the processing results of laser point cloud big data, the line features of building space are extracted based on the improved Mean-shift method, and the continuous broken lines in the point cloud data of building space are extracted by using the double radius threshold line tracing method. According to the feature extraction results of point cloud data in building space, the visual reconstruction of building space is completed through the process of translation matching and space matching. The experimental results show that this method can realize the visual reconstruction of architectural space, and the average reconstruction accuracy is higher than that of 97 %, and the reconstruction completion and smoothness are higher than 95 %.

A Deep-Learning-Based Method for Extracting an Arbitrary Number of Individual Power Lines from UAV-Mounted Laser Scanning Point Clouds

In recent years, laser scanners integrated with Unmanned Aerial Vehicles (UAVs) have exhibited great potential in conducting power line inspections in harsh environments. The point clouds collected for power line inspections have numerous advantages over remote image data. However, point cloud-based individual power line extraction, which is a crucial technology required for power line inspections, still poses several challenges such as massive 3D points, imbalanced category points, etc. Moreover, in various power line scenarios, previous studies often require manual setup and careful adjustment of different thresholds to separate different power lines, which is inefficient for practical applications. To handle these challenges, in this paper, we propose a multi-branch network to automatically extract an arbitrary number of individual power lines from point clouds collected by UAV-based laser scanners. Specifically, to handle the massive 3D point clouds in complex outdoor scenarios, we propose to leverage deep neural network for efficient and rapid feature extraction in large-scale point clouds. To mitigate imbalanced data quantities across different categories, we propose to design a weighted cross-entropy loss function to measure the varying importance of each category. To achieve the effective extraction of an arbitrary number of power lines, we propose leveraging a loss function to learn the discriminative features that can differentiate the points belonging to different power lines. Once the discriminative features are learned, the Mean Shift method can distinguish the individual power lines by clustering without supervision. The evaluations are executed on two datasets, which are acquired at different locations with UAV-mounted laser scanners. The proposed method has been thoroughly tested and evaluated, and the results and discussions confirm its outstanding ability to extract an arbitrary number of individual power lines in point clouds.

Detection of Apparent Defects in HPLC/Dual Mode Portable Sorting Device Based on Deep Learning and Image Processing

Abstract As the growth of automated detection technology, traditional manual detection has gradually been replaced. To improve the effectiveness of defect detection, a HPLC/dual mode portable sorting device with deep learning and image processing is raised for apparent defect detection. The product image is segmented using flood filled mean shift method, and defect detection is achieved using Yolo v3 algorithm. An LME2918 chip is the main component of the communication module in the portable device. Based on experimental data, image segmentation accuracy can range from 80% to 100%, and image matching accuracy can range from 85% to 95%. Under the power line carrier mode, the average success rate of apparent defect detection in portable sorting devices can reach 85%, and the change in success rate is not significant, indicating that the detection is relatively stable; Under wireless communication mode, the average success rate can reach 83%, and the change in success rate is relatively obvious, because wireless communication is easily affected by the external environment. The experimental data shows that the defect detection effect of the HPLC/dual mode portable sorting device based on deep learning and image processing meets the design requirements.

Identification of the selected soil bacteria genera based on their geometric and dispersion features.

The visual analysis of microscopic images is often used for soil bacteria recognition in microbiology. Such task can be automated with the aid of machine learning and digital image processing techniques. The best results for soil microorganism identification usually rely on extracting features based on color. However, accommodating in the latter an extra impact of lighting conditions or sample's preparation on classification accuracy is often omitted. In contrast, this research examines features which are insensitive to the above two factors by focusing rather on bacteria shape and their specific group dispersion. In doing so, the calculation of layout features resorts to k-means and mean shift methods. Additionally, the dependencies between specific distances determined from bacteria cells and the curvature of interpolated bacteria boundary are computed to extract vital geometric shape information. The proposed bacteria recognition tool involves testing four different classification methods for which the parameters are iteratively adjusted. The results obtained here for five selected soil bacteria genera: Enterobacter, Rhizobium, Pantoea, Bradyrhizobium and Pseudomonas reach 85.14% classification accuracy upon combining both geometric and dispersion features. The latter forms a promising result as a substitutive tool for color-based feature classification.

NMSFRA: Heterogeneous routing protocol for balanced energy consumption in mobile wireless sensor network

Wireless sensor networks are widely used in various fields and consist of a large number of sensor nodes that are powered by batteries. As a result, energy efficiency is crucial for the effective operation of these networks. With the advancement of cross technology communication (CTC), the presence of heterogeneous wireless nodes, especially mobile nodes, in wireless sensor networks has increased, making traditional routing protocols for homogeneous wireless sensor networks inapplicable. To overcome the above issues, this paper proposes a heterogeneous wireless sensor network routing protocol for a balanced energy consumption, NMSFRA (NGO based mean shift fuzzy clustering routing algorithm). In order to mitigate the impact of uneven cluster distribution and unstable network links due to the mobility of nodes, the proposed protocol first divides the network into balanced clusters using the Mean Shift (MS) method. The Mamdani fuzzy inference system is then used to take node mobility into account. The protocol dynamically selects different parameters as inputs to the fuzzy logic, enabling it to choose cluster heads (CHs) based on the overall location of the clusters. Additionally, to extend the network lifetime, the protocol generates multi-hop forwarding paths using the Northern Goshawk Optimization (NGO) method. Finally, the paper presents experimental simulations carried out on three different network topologies and compares the results with existing protocols such as LEACH, EETPF, E-FUCA, and DE-SEP. The results show that the NMSFRA protocol extends the survival period of nodes by 235 rounds, 240 rounds, 364 rounds, and 190 rounds, on average, compared to the LEACH, EETPF, E-FUCA, and DE-SEP protocols, respectively. The network throughput also increased by 154.92%, 48.74%, 84.77%, and 106.77%, respectively. The NMSFRA protocol also showed better performance in terms of total residual energy, FND, HND, and average speed difference between nodes in the first 5 rounds. Therefore, the proposed NMSFRA protocol demonstrates superior performance in terms of network lifetime, energy utilization, throughput, and network stability.

Segmentation of High‐Resolution Remote Sensing Images Using the Gabor Texture Feature‐Based Mean Shift Method

High‐resolution remote sensing images (HRRSIs) play an important role in the construction and development of society with their rich and detailed information. In the process of remote sensing image segmentation, the conventional method of mean shift usually involves spatial features and spectral features to preserve edge information and reduce the effect of noise. However, this traditional method often misses some important information in HRRSIs processing and reduces the segmentation accuracy. Thus, we proposed an improved mean shift image segmentation method based on Gabor texture features to make full use of the delicate information in HRRSIs. Specifically, the multiscale and multidirectional Gabor filters were employed in this study to extract pixel‐to‐pixel features from the QuickBird and GeoEye‐1 images, and the local variance was used to determine the optimal texture bandwidth. Two parameters, global segmentation (GS) quality index and error rate (ER), were adopted to evaluate the segmentation quality of these two HRRSIs. Results showed that the proposed approach in this study outperforms the traditional mean shift method on both remote sensing datasets. Compared with the traditional method, the improved mean shift approach corresponded to higher GS values and smaller ER values. This study confirms that the improved mean shift approach based on Gabor filtering has a potential for HRRSIs segmentation.

Sconce: a cosmic web finder for spherical and conic geometries.

The latticework structure known as the cosmic web provides a valuable insight into the assembly history of large-scale structures. Despite the variety of methods to identify the cosmic web structures, they mostly rely on the assumption that galaxies are embedded in a Euclidean geometric space. Here, we present a novel cosmic web identifier called sconce (Spherical and CONic Cosmic wEb finder) that inherently considers the 2D (RA, DEC) spherical or the 3D (RA, DEC, z) conic geometry. The proposed algorithms in sconce generalize the well-known subspace constrained mean shift (scms) method and primarily address the predominant filament detection problem. They are intrinsic to the spherical/conic geometry and invariant to data rotations. We further test the efficacy of our method with an artificial cross-shaped filament example and apply it to the SDSS galaxy catalogue, revealing that the 2D spherical version of our algorithms is robust even in regions of high declination. Finally, using N-body simulations from Illustris, we show that the 3D conic version of our algorithms is more robust in detecting filaments than the standard scms method under the redshift distortions caused by the peculiar velocities of haloes. Our cosmic web finder is packaged in python as sconce-scms and has been made publicly available.

Shadow Compensation from UAV Images Based on Texture-Preserving Local Color Transfer

The lack of color information and texture information in the shadow region seriously affect the recognition and interpretation of remote sensing image information. The commonly used methods focus on the restoration of texture information, but it is often easy to overcompensate, resulting in color distortion of the shadow region. However, some methods only ensure accurate correction of color information, and tend to cause texture blurring. In order to not lose the texture information and to accurately compensate the color information in the shadow region of the image, we propose a shadow compensation method from UAV images based on texture-preserving local color transfer in this paper. Firstly, homogeneous regions are extracted from UAV images and homogeneous subregion segmentation is performed on the basis of homogeneous regions using the mean shift method. Secondly, in combination with the shadow mask, each shadow subregion is matched with the corresponding non-shadow subregion based on its texture features and spatial distance. Then, the matched non-shadow subregion is used as the reference region, and the color transfer based on preserving texture is performed on the shadow subregion. Finally, pixel-by-pixel width shadow compensation is applied to the penumbra region. The results of the qualitative and quantitative analysis validate the accuracy and effectiveness of the proposed methodology to compensate for the color and texture details of the shadow regions.

Research and Application of Computer Collaborative Spectrum Sensing Guided by Radio Technology

Currently, due to the proliferation of smart appliances and the rapid development of wireless communication technologies, there is a huge shortage of spectrum resources. Cognitive radio utilizes technologies that dramatically increase the utilization of free frequency bands, detecting the environment in real time and maximizing the reallocation of spectrum resources by sensing and reallocating the licensed free frequency bands. The frequency and accuracy required to determine the free bands is an important basis for determining the spectrum awareness in all aspects of a sensing radio system. Therefore, for the more complex network environment in reality and the intrusion of malicious users, this paper uses Fast K-medoids and Mean-shift methods for information fusion and an iterative approach to overcome malicious use to fuse data and finally obtain realistic feature statistics. The results show that the collaborative identification strategy proposed in this paper has good results.

A high-resolution map of coastal vegetation for two Arctic Alaskan parklands: An object-oriented approach with point training data.

Bering Land Bridge National Preserve and Cape Krusenstern National Monument in northwest Alaska have approximately 1600 km of predominantly soft-sediment coastlines along the Chukchi Sea, a shallow bay of the Arctic Ocean. Over the past decade, marine vessel traffic through the Bering Strait has grown exponentially to take advantage of new ice-free summer shipping routes, increasing the risk of oil spills in these fragile ecosystems. We present a high-resolution coastal vegetation map to serve as a baseline for potential spill response, restoration, and change detection. We segmented 663 km2 of high-resolution multispectral satellite images by the mean-shift method and collected 40 spectral, topographic and spatial variables per segment. The segments were classified using photo-interpreted points as training data, and verified with field based plots. Digitizing points, rather than polygons, and intersecting them with the segmentation allows for rapid collection of training data. We classified the map segments using Random Forest because of its high accuracy, computational speed, and ability to incorporate non-normal, high-dimensional data. We found creating separate classification models by each satellite scene gave highly similar results to models combining the entire study area, and that reducing the number of variables had little impact on accuracy. A unified, study area-wide Random Forest model for both parklands produced the highest accuracy of various models attempted. We mapped 18 distinct classes, with an out-of-bag error of 11.6%, resulting in an improvement to the past per-pixel classification of this coast, and in higher spatial and vegetation classification resolution. The resulting map demonstrates the utility of our point-based method and provides baseline data for incident preparedness and change detection. Elevation is highly correlated with the ordination of the vegetation types, and was the most important variable in all tested classification models. The vegetation classification brings together the largest amount of vegetation data for the Chukchi Sea coast yet documented.

Research on Digital Steganography and Image Synthesis Model Based on Improved Wavelet Neural Network.

Network compression coding technology is a research hotspot in the field of digital steganography and image synthesis. How to improve image quality while achieving short compression time is a problem currently faced. Based on the improved wavelet neural network theory, this paper constructs a digital steganography and image synthesis model. The model first tracks the contour of the digit to be recognized, then equalizes and resamples the contour to make it translation-invariant and scaling-invariant, and then uses multi-wavelet neural network clusters to stretch the contour shell to obtain orders of magnitude multi-resolution and its average, and finally, these shell coefficients are fed into a feedforward neural network cluster to identify this handwritten digit, solving the problem of multi-resolution decomposition of contour shells while having a high sampling rate. In the simulation process, the classification model that a single pixel is a text/non-text pixel is trained on the original gray value of the gray pixel and its neighboring pixels, and the classified text pixels are connected to a text area through an adaptive MeanShift method. The experimental results show that it is feasible to use multi-wavelet features for handwritten digit recognition. The model combines the neural network and the genetic algorithm, making full use of the advantages of both, so that the new algorithm has the learning ability and robustness of the neural network. The compression ratio after compression by ordinary wavelet coding, PSNR, MSE, and compression time are 8.4, 25 dB, 210, and 7 s, respectively. The values are 11.7, 24 dB, 207, and 11 s, respectively. At the same time, the peak signal-to-noise ratio is higher and the mean square error is lower, that is, the compression quality is better, and the accuracy of digital steganography and image synthesis is effectively improved.

Study of Clustering Algorithms in Object Tracking and Image Segmentation

Mean Shift is a kind of clustering algorithm, which is mostly used for target tracking, image segmentation, etc. In order to solve the problem that image information is not effectively utilized because of unclear traffic video images and random jitter between image sequences, this paper has studied how to achieve stability of traffic video images and proposed an improved Mean Shift algorithm about how to conduct object centroid registration in compensation for the deviations in space localization and, on this basis, how to select kernel window width to eliminate the errors in scale positioning. This algorithm gets the tracking effect and computation analysis of improved Mean Shift from the perspective of applications, removes or relieves the impact motion has on imaging, improves the quality of the video image information obtained, automatically adjusts the size of window according to the scale changes of moving object in the image, and effectively enhances the stability and real time of object tracking. Besides, in the postprocessing stage, the superpixel based on the Mean Shift algorithm is applied to further optimize the segmentation result; it is a popular mode-seeking clustering algorithm, which makes Mean Shift method ideal on low-dimensional applications such as image segmentation. Finally, we show promising results for remote sensing image segmentation.

A Study on Region-Based Image Segmentation Techniques

This paper addresses the comparison of two popular region-based image segmentation techniques namely the Watershed method and the Mean-shift method. The watershed method is a mathematical morphology based image segmentation approach while Mean-shift method is a data-clustering method that searches for the local maximal density points and then groups all the data to the clusters defined by these maximal density points. Here the efficiency of the both segmentation techniques is presented

Color Image Segmentation by Utilizing Coarse-to-Fine Strategy

Image segmentation is an important processing technology, which is the basis of image recognition and has been widely used in many fields. In this paper, we propose a method, termed coarse-to-fine strategy-based image segmentation (CSIS), for color image segmentation. The basic idea is to segment an image by three phases: (1) the original image is first segmented into several distinct regions by using the mean shift method; (2) the segmented regions are converted to a weighted region adjacency graph (RAG), and a new graph cut method, called multi-cut algorithm, is proposed to partition the RAG into multiple regions; (3) a one-step Chan–Vese algorithm is applied to smooth the boundaries of the segmented objectives. In each of the last two phases, a method is applied to refine the result obtained in the previous phase. By carefully balancing the efforts used in each phase, CSIS could segment color images both efficiently and effectively. These advantages are demonstrated by applying the proposed method to a variety of test instances, and the statistical results also show that it is comparable with some state-of-the-art methods.

A Novel Mean-Shift Algorithm for Data Clustering

We propose a novel Mean-Shift method for data clustering, called Robust Mean-Shift (RMS). A new update equation for point iterates is proposed, mixing the ones of the standard Mean-Shift (MS) and the Blurring Mean-Shift (BMS). Despite its simplicity, the proposed method has not been studied so far. RMS can be set up in both a kernel-based and a nearest-neighbor (NN)-based fashion. Since the update rule of RMS is closer to BMS, the convergence of point iterates is conjectured based on the Chen’s BMS convergence theorem. Experimental results on synthetic and real datasets show that RMS in several cases outperforms MS and BMS in the clustering task. In addition, RMS exhibits larger attraction basins than MS and BMS for identical parametrization; consequently, its kernel variant requires a lower aperture of the kernel function, and its NN variant a lower number of nearest neighbors compared to MS or BMS, to achieve optimal clustering results. In addition, the NN version of RMS does not need to specify a convergence threshold to stop the iterations, contrarily to the NN-BMS algorithm.