An Improved Optical Flow Algorithm Based on Mask-R-CNN and K-Means for Velocity Calculation

Peng, Yahui; Liu, Xiaochen; Shen, Chong; Huang, Haoqian; Zhao, Deli; Cao, Huiliang; Guo, Xiaoting

doi:10.3390/app9142808

Cited by 14 publications

(8 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Considering the fact that the maceral components within each group mostly are not adjacent to each other and the gray scale values are the major difference between maceral groups, we adopt the gray scale values of each pixel as the features. K-means clustering, one of the most favorable clustering techniques, is utilized for its simplicity and computational efficiency [21,22].…”

Section: Image Segmentation Based On Two-level Clusteringmentioning

confidence: 99%

Intelligent Identification of Maceral Components of Coal Based on Image Segmentation and Classification

et al. 2019

View full text Add to dashboard Cite

Featured Application: Maceral Analysis; Coal Processing.Abstract: An intelligent analytical technique which is able to accurately identify maceral components is highly desired in the fields of mining and geology. However, currently available methods based on fixed-size window neglect the shape information, and thus do not work in identifying maceral composition from one entire photomicrograph. To address these concerns, we propose a novel Maceral Identification strategy based on image Segmentation and Classification (MISC). Considering the complex and heterogeneous nature of coal, a two-level coarse-to-fine clustering method based on K-means is employed to divide microscopic images into a sequence of regions with similar attributes (i.e., binder, vitrinite, liptinite and inertinite). Furthermore, comprehensive features along with random forest are utilized to automatically classify binder and seven types of maceral components, including vitrinite, fusinite, semifusinite, cutinite, sporinite, inertodetrinite and micrinite. Evaluations on 39 microscopic images show that the proposed method achieves the state-of-the-art accuracy of 90.44% and serves as the baseline for future research on maceral analysis. In addition, to support the decisions of petrologists during maceral analysis, we developed a standalone software, which is freely available at https:/github.com/GuyooGu/MISC-Master.

show abstract

Section: Image Segmentation Based On Two-level Clusteringmentioning

confidence: 99%

Intelligent Identification of Maceral Components of Coal Based on Image Segmentation and Classification

et al. 2019

View full text Add to dashboard Cite

show abstract

“…According to the horn&schunck optical ow algorithm of gradient, an improved optical ow method of Gaussian pyramid is proposed in the literature. This method combines the moving target segmentation with the maximum interclass variance and mathematical morphology operation, and carries out HS calculation, which effectively reduces the number of iterations of optical ow calculation, improves the detection e ciency of the algorithm, and basically meets the requirements of real-time [7][8]. Aiming at the problem of slow convergence of Gaussian mixture model, an improved algorithm is proposed to improve the convergence rate without affecting the stability of the model.…”

Section: Related Workmentioning

confidence: 99%

Application of adaptive moving object detection and edge calculation in police physical education training app course

Zhen

2023

Soft Comput

View full text Add to dashboard Cite

Nowadays, the theoretical knowledge and computer skills of digital signal processing are increasing year by year, the mode of combining moving target detection industry and edge computing gradually occupies an important position in the research process of processing and pattern recognition in various application elds.For moving target detection and edge algorithm, this paper proposes a technical method based on Gaussian mixture background model and ltering, which combines moving target detection and edge calculation. This method changes the learning rate mechanism of the traditional weight in the background model updating, so that the background model can be continuously effective. Then the motion information and centroid position information of the detected moving target are taken as the initial input data of the lter, so that the lter can start tracking the target. Finally, by analyzing and understanding the development history of Chinese police sports, analyzing the current situation of police sports teaching and training in Chinese public security colleges, we nd the problems existing in police sports teaching and training in public security colleges, and use target detection and edge algorithm. Carry out actual calculation in the police sports teaching app, adjust the curriculum objective system of police sports teaching, update the teaching content, reform the teaching methods, reposition the police sports teaching curriculum, build a high-quality teaching team, fully improve the theoretical system of police sports quality education in public security colleges, and cultivate a large number of compound and applied public security talents with strong comprehensive quality and ability.

show abstract

“…K-means clustering is a clustering algorithm whose main objective is to group elements or similar data points in a cluster [2,15,16] where K represents the number of clusters or groups. The algorithm then runs iteratively to assign each data point to one of the k groups based on the functionality already provided.…”

Section: K-means Clusteringmentioning

confidence: 99%

Image Segmentation Approach Based on Hybridization Between K-Means and Mask R-CNN

Moussaoui

Benslimane

Akkad

2021

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

In this article, we will introduce a hybrid method based on the combination of two image segmentation techniques. The first method adopted is the k-means algorithm which is an unsupervised machine learning technique used to group data points, the second is the Mask R-CNN which is a neural network architecture which combines two sub-problems: object detection and semantic segmentation. The main objective of this study is to approve the segmentation of the image using k-means. The first step is to apply Mask R-CNN on our original image to detect the objects that are present in the image, then, we will apply k-means clustering to have better segmentation. For our approach we used a set of metrics to evaluate our proposed approach such as the mean square error (MSE); the peak signal to noise ratio (PSNR) and many other measures such as the difference between two images. These measurements have shown satisfactory results and a performance of the proposed method.

show abstract

An Improved Optical Flow Algorithm Based on Mask-R-CNN and K-Means for Velocity Calculation

Cited by 14 publications

References 38 publications

Intelligent Identification of Maceral Components of Coal Based on Image Segmentation and Classification

Intelligent Identification of Maceral Components of Coal Based on Image Segmentation and Classification

Application of adaptive moving object detection and edge calculation in police physical education training app course

Image Segmentation Approach Based on Hybridization Between K-Means and Mask R-CNN

Contact Info

Product

Resources

About