scholarly journals Real-Time Depth-Based Hand Detection and Tracking

2014 ◽  
Vol 2014 ◽  
pp. 1-17 ◽  
Author(s):  
Sung-Il Joo ◽  
Sun-Hee Weon ◽  
Hyung-Il Choi
Keyword(s):  
Real Time ◽  
Mean Shift ◽  
Region Growing ◽  
Tracking Accuracy ◽  
Hand Detection ◽  
Seed Point ◽  
Mean Shift Algorithm ◽  
Detection And Tracking ◽  
Cascade Structure ◽  
Hand Region

This paper illustrates the hand detection and tracking method that operates in real time on depth data. To detect a hand region, we propose the classifier that combines a boosting and a cascade structure. The classifier uses the features of depth-difference at the stage of detection as well as learning. The features of each candidate segment are to be computed by subtracting the averages of depth values of subblocks from the central depth value of the segment. The features are selectively employed according to their discriminating power when constructing the classifier. To predict a hand region in a successive frame, a seed point in the next frame is to be determined. Starting from the seed point, a region growing scheme is applied to obtain a hand region. To determine the central point of a hand, we propose the so-called Depth Adaptive Mean Shift algorithm. DAM-Shift is a variant of CAM-Shift (Bradski, 1998), where the size of the search disk varies according to the depth of a hand. We have evaluated the proposed hand detection and tracking algorithm by comparing it against the existing AdaBoost (Friedman et al., 2000) qualitatively and quantitatively. We have analyzed the tracking accuracy through performance tests in various situations.

scholarly journals Hand Detection Using Cascade of Softmax Classifiers

2018 ◽  
Vol 2018 ◽  
pp. 1-11
Author(s):  
Yan-Guo Zhao ◽  
Feng Zheng ◽  
Zhan Song
Keyword(s):  
Logistic Regression ◽  
Real Time ◽  
Sliding Window ◽  
Detection Accuracy ◽  
Hand Posture ◽  
Hand Detection ◽  
Cascade Structure ◽  
Real Time Applications ◽  
Binary Classifiers ◽  
Background Pattern

Sliding-window based multiclass hand posture detections are often performed by detecting postures of each predefined category using an independent detector, which makes it lack efficiency and results in high postures confusion rates in real-time applications. To tackle such problems, in this work, an efficient cascade detector that integrates multiple softmax-based binary (SftB) models and a softmax-based multiclass (SftM) model is investigated to perform multiclass posture detection in parallel. The SftB models are used to distinguish the predefined postures from the background regions, and the SftM model is applied to discriminate among all the predefined hand posture categories. Another usage of the cascade structure is that it could effectively decompose the complexity of background pattern space and therefore improve the detection accuracy. In addition, to balance the detection accuracy and efficiency, the HOG features of increasing resolutions will be adopted by classifiers of increasing stage-levels in the cascade structure. The experiments are implemented under various scenarios with complicated background and challenging lightings. Results show the superiority of the proposed SftB classifiers over the traditional binary classifiers such as logistic regression, as well as the accuracy and efficiency improvements brought by the softmax-based cascade architecture compared with the noncascade multiclass softmax detectors.

Tracking algorithm of snowboard target in intelligent system

2020 ◽  
pp. 1-9
Author(s):  
Zhipeng Li ◽  
Xiaolan Li ◽  
Ming Shi ◽  
Wenli Song ◽  
Guowei Zhao ◽  
...  
Keyword(s):  
Target Tracking ◽  
High Speed ◽  
Intelligent System ◽  
Mean Shift ◽  
Target Position ◽  
Tracking Algorithm ◽  
Tracking Accuracy ◽  
Mean Shift Algorithm ◽  
Target Track ◽  
The Mean

Snowboarding is a kind of sport that takes snowboarding as a tool, swivels and glides rapidly on the specified slope line, and completes all kinds of difficult actions in the air. Because the sport is in the state of high-speed movement, it is difficult to direct guidance during the sport, which is not conducive to athletes to find problems and correct them, so it is necessary to track the target track of snowboarding. The target tracking algorithm is the main solution to this task, but there are many problems in the existing target tracking algorithm that have not been solved, especially the target tracking accuracy in complex scenes is insufficient. Therefore, based on the advantages of the mean shift algorithm and Kalman algorithm, this paper proposes a better tracking algorithm for snowboard moving targets. In the method designed in this paper, in order to solve the problem, a multi-algorithm fusion target tracking algorithm is proposed. Firstly, the SIFT feature algorithm is used for rough matching to determine the fuzzy position of the target. Then, the good performance of the mean shift algorithm is used to further match the target position and determine the exact position of the target. Finally, the Kalman filtering algorithm is used to further improve the target tracking algorithm to solve the template trajectory prediction under occlusion and achieve the target trajectory tracking algorithm design of snowboarding.

Study of Missile Radiator Predicting and Tracking Technology on the Basis of Kalman Filter and Mean-Shift Algorithm

2011 ◽  
Vol 383-390 ◽  
pp. 1584-1589
Author(s):  
Zhen Hui Xu ◽  
Bao Quan Mao ◽  
Li Xu ◽  
Jun Yan Zhao
Keyword(s):  
Kalman Filter ◽  
Real Time ◽  
Kalman Filtering ◽  
Mean Shift ◽  
Moving Target ◽  
Tracking Method ◽  
The Real ◽  
Mean Shift Algorithm ◽  
Compensation Function ◽  
Tracking Technology

In order to improve the real-time character of missile radiator tracking and solve the predicting tracking problem when missile radiator shortly shelter or missing, it introduces moving target predicting and tracking technology. According to the predicting and tracking method, it proposes three predicting and tracking overall schemes of missile radiator based on Kalman filtering and improved Mean-Shift algorithm. Also it compares the real-time character of three kinds of schemes. According to the trajectory character of missile radiator, it constructs Kalman filter. The experiment results indicate that by using Kalman filtering technology, there are improvements in real-time character and shortly shelter or missing problem can be solved well. It plays a certain compensation function to the whole system.

Gesture Tracking and Locating Algorithm Based on Federated Tracking Filter

2019 ◽  
Vol 141 (12) ◽  
Author(s):  
Lei Yu ◽  
Junyi Hou ◽  
Shumin Fei
Keyword(s):  
Mean Shift ◽  
Experimental Simulation ◽  
Human Hand ◽  
Sampling Point ◽  
Tracking Accuracy ◽  
Mean Shift Algorithm ◽  
Gesture Tracking ◽  
Tracking Time ◽  
Tracking Filter ◽  
The Impact

Abstract In this paper, a joint gesture tracking method combining particle filter and mean shift algorithm is proposed to improve the accuracy and robustness of the system. During the slow movement of the human hand, the average movement of the particles is first used to cause most of the particles to drift into the gesture area. In the case where the movement of the human hand is faster or there is occlusion, when the mean shift of the particle is performed, if the region of the gesture is not detected, the particle will return to the state before the drift, and then the next frame is processed. The method can directly calculate the position of the gesture based on the particles used for subsequent testing, and can save the tracking time of the algorithm. Through experimental simulation, compared with the Cam-shift algorithm, when the sampling point of the joint tracking algorithm proposed in this paper is 200, the tracking accuracy is improved to 95.2%. Compared with 90.6% of the Cam-shift algorithm, the tracking time is reduced from 83.7 ms to 25.8 ms. Therefore, the proposed algorithm can greatly improve the tracking accuracy and real-time, and can also effectively reduce the impact of complex environments on the tracking effect.

scholarly journals A High-Precision, Real-Time, and Robust Indoor Visible Light Positioning Method Based on Mean Shift Algorithm and Unscented Kalman Filter

Sensors ◽  
2019 ◽  
Vol 19 (5) ◽  
pp. 1094 ◽  
Author(s):  
Zekun Xie ◽  
Weipeng Guan ◽  
Jieheng Zheng ◽  
Xinjie Zhang ◽  
Shihuan Chen ◽  
...  
Keyword(s):  
Kalman Filter ◽  
Visible Light ◽  
Real Time ◽  
High Speed ◽  
Unscented Kalman Filter ◽  
Light Emitting Diode ◽  
Mean Shift ◽  
Indoor Navigation ◽  
Positioning Accuracy ◽  
Mean Shift Algorithm

Visible light positioning (VLP) is a promising technology for indoor navigation. However, most studies of VLP systems nowadays only focus on positioning accuracy, whereas robustness and real-time ability are often overlooked, which are all indispensable in actual VLP situations. Thus, we propose a novel VLP method based on mean shift (MS) algorithm and unscented Kalman filter (UKF) using image sensors as the positioning terminal and a Light Emitting Diode (LED) as the transmitting terminal. The main part of our VLP method is the MS algorithm, realizing high positioning accuracy with good robustness. Besides, UKF equips the mean shift algorithm with the capacity to track high-speed targets and improves the positioning accuracy when the LED is shielded. Moreover, a LED-ID (the identification of the LED) recognition algorithm proposed in our previous work was utilized to locate the LED in the initial frame, which also initialized MS and UKF. Furthermore, experiments showed that the positioning accuracy of our VLP algorithm was 0.42 cm, and the average processing time per frame was 24.93 ms. Also, even when half of the LED was shielded, the accuracy was maintained at 1.41 cm. All these data demonstrate that our proposed algorithm has excellent accuracy, strong robustness, and good real-time ability.

Detection of Hands for Hand-Controlled Skyfall Game in Real Time Using CNN

2020 ◽  
Vol 10 (2) ◽  
pp. 15-25
Author(s):  
Neha B. ◽  
Naveen V. ◽  
Angelin Gladston
Keyword(s):  
Real Time ◽  
Detection Methods ◽  
Input Device ◽  
Single Shot ◽  
Hand Detection ◽  
Web Based ◽  
Fast Detection ◽  
Detection And Tracking ◽  
Average Accuracy ◽  
Direct Use

With human-computer interaction technology evolving, direct use of the hand as an input device is of wide attraction. Recently, object detection methods using CNN models have significantly improved the accuracy of hand detection. This paper focuses on creating a hand-controlled web-based skyfall game by building a real time hand detection using CNN-based technique. A CNN network, which uses a MobileNet as the feature extractor along with the single shot detector framework, is used to achieve a robust and fast detection of hand location and tracking. Along with detection and tracking of hand, skyfall game has been designed to play using hand in real time with tensor flow framework. This way of designing the game where hand is used as input to control the paddle of skyfall game improved the player interaction and interest towards playing the game. This model of CNN network used egohands dataset for detecting and tracking the hands in real time and produced an average accuracy of 0.9 for open hands and 0.6 for closed hands which in turn improved player and game interactions.

scholarly journals Implementation of Camshift Target Tracking Algorithm Based on Hybrid Filtering and Multifeature Fusion

2020 ◽  
Vol 2020 ◽  
pp. 1-13
Author(s):  
Sijie Du ◽  
Hongxin Xu ◽  
Tianping Li
Keyword(s):  
Kalman Filter ◽  
Particle Filter ◽  
Low Cost ◽  
Mean Shift ◽  
Video Tracking ◽  
Experimental Result ◽  
Tracking Accuracy ◽  
Mean Shift Algorithm ◽  
Positioning Errors ◽  
Tracking Process

In recent years, the Mean shift algorithm has extensive applications in the field of video tracking. It has some advantages of low cost, small memory, and good tracking effect. However, there are some shortcomings in the existing algorithm; for example, it cannot produce adaptive changes as the target size changes. And when there are similar objects, it is prone to target positioning errors and tracking failures caused by occlusion. In this paper, an improved method of continuous adaptive change Mean shift (Camshift) for high-precision positioning and tracking is proposed. The traditional Camshift method only uses hue components in HSV to extract features. This paper uses the combination of H and S components in HSV space to build a two-dimensional color feature histogram and with the image’s LBP feature histogram to increase tracking accuracy. Meanwhile, for the sake of target occlusion and nonlinear changes in the tracking process, this paper introduces a Gaussian-Hermit particle filter that is updated by the Kalman filter. Experimental result demonstrates that the real-time performance of the proposal in this paper is better than Mean shift, Camshift, simple particle filter, and Kalman filter.

A Novel Mesh Segmentation Method for Conventional Engineering Parts in Reverse Engineering

2012 ◽  
Vol 215-216 ◽  
pp. 660-663
Author(s):  
Peng Xin Liu ◽  
Xuan Liu ◽  
Ping Liu
Keyword(s):  
Reverse Engineering ◽  
Mean Shift ◽  
Region Growing ◽  
Mesh Segmentation ◽  
Image Process ◽  
Segmentation Method ◽  
Mean Shift Algorithm ◽  
Noise Data ◽  
Normal Vectors ◽  
Scan Data

In this paper a step segmentation method for conventional engineering parts which usually comprise by simple analytic surfaces and are important investigation objects of reverse engineering is proposed. In the first segmentation fourteen directions are chosen initially, and a k-mean algorithm according to the normal vectors is used, then the surface is divided to form patches by a region-growing scheme. In order to identify simple analytic surface, curvedness of a patch is set as the criterion by which the surface merged into near constant curvedness region. Especially a novel mean shift algorithm is adopted in this method, that a powerful technique for clustering in image process, and is extended to normal filtering while preserving the features to increase robustness of the method. Experimental evaluations using scan data or noise data demonstrate the efficiency of the proposed method.

Research on target tracking algorithm based on mean shift with adaptive bandwidth

2021 ◽  
pp. 1-15
Author(s):  
Ming Han ◽  
Jingqin Wang ◽  
Jingtao Wang ◽  
Junying Meng ◽  
Ying Cheng
Keyword(s):  
Target Tracking ◽  
Real Time ◽  
Kernel Function ◽  
Mean Shift ◽  
Tracking Algorithm ◽  
Distance Constraint ◽  
Mean Shift Algorithm ◽  
Signed Distance ◽  
Adaptive Bandwidth ◽  
Anisotropic Kernel

The traditional mean shift algorithm used fixed kernels or symmetric kernel function, which will cause the target tracking lost or failure. The target tracking algorithm based on mean shift with adaptive bandwidth was proposed. Firstly, the signed distance constraint function was introduced to produce the anisotropic kernel function based on signed distance kernel function. This anisotropic kernel function satisfies that the value of the region function outside the target is zero, which provides accurate tracking window for the target tracking. Secondly, calculate the mean shift window center of anisotropic kernel function template, the theory basis is the sum of vector weights from the sample point in the tracking window to the center point is zero. Thirdly, anisotropic kernel function templates adaptive update implementation by similarity threshold to limit the change of the template between two sequential pictures, so as to realize real-time precise tracking. Finally, the contrast experimental results show that our algorithm has good accuracy and high real time.

Sign in / Sign up

Export Citation Format

Share Document