ArticlePDF Available

Human Computer Interaction using Hand Gesture Recognition

Authors:
Selvarathi Chelliah at M.Kumarasamy College of Engineering
Selvarathi Chelliah
Download full-text PDF
Read full-text
Download citation

Figures

Figures - uploaded by Selvarathi Chelliah
Author content
All figure content in this area was uploaded by Selvarathi Chelliah
Content may be subject to copyright.
Content uploaded by Selvarathi Chelliah
Author content
All content in this area was uploaded by Selvarathi Chelliah on Aug 18, 2022
Content may be subject to copyright.
C.Selvarathi et al., International Journal of Advanced Trends in Computer Science and Engineering, 9(2), March - April 2020, 1600 – 1603
1600
Human Computer Interaction using Hand Gesture Recognition
C.Selvarathi1, P.Indu2 , B. Kavyadharshini3 , S. Logesh kumar4 , R. Mohamed yasher5
1 Department of CSE, M.Kumarasamy College of Engineering, India,selvarathic.cse@mkce.ac.in
2Department of CSE, M.Kumarasamy College of Engineering, India, indu270699@gmail.com
3Department of CSE, M.Kumarasamy College of Engineering, India, kavyadharshinibg@gmail.com
4Department of CSE, M.Kumarasamy College of Engineering, India, slogesh1999@gmail.com
5Department of CSE, M.Kumarasamy College of Engineering, India, rmyasher1998@gmail.com
ABSTRACT
Hand gesture recognition provides an interaction between
human and PCs. Its applications extend from therapeutic
recovery to purchaser gadgets control (for example cell
phone). To identify hand gestures, different methods are
used. Signals a non-verbal type of correspondence gives
the HCI interface. Ongoing vision-based hand motion
acknowledgment is viewed as increasingly more possible
for HCI with the assistance of most recent advances in
the field of PC vision. This venture manages conversation
of different procedures, techniques and calculations
identified with the motion acknowledgment. The hand
motion is the most simple and common method for
correspondence. Hand motion acknowledgment has the
different favorable circumstances of ready to speak with
the Technology through fundamental communication via
gestures. The motion will ready to diminish the utilization
of most noticeable equipment gadgets which are utilized
to control the exercises of the PC.
Key words: Hand gestures, Foreground and background
substraction, finger count detection.
1. INTRODUCTION
IOT plays a vital role in system administration and other
related framework activities. The framework has two
significant focal points and three stages. Each stage is
capsuled to another. The edges of hand which is
recognized as an extralayer, used in different applications.
Edge location is one of the most normally utilized tasks in
picture examination, and there are likely more
calculations in the writing for improving and identifying
edges than some other single subject. The explanation
behind this is edges structure the framework of an
article.Anedge is the limit between an article and the
foundation, and demonstrates the limit between covering
objects. This implies if the edges in a picture can be
recognized precisely, the entirety of the items can be
found and fundamental properties, for example, region,
edge, and shape can be estimated.
2. LITERATURE SURVEY
The investigation is devoted to the production of program
and innovative automated applications that will empower
significantly improve things in regards to language,
especially inside the cases once the same correspondence
is out there. Albeit a static hand signal might be any
feasible stance of a human's hand, regularly exclusively a
confined arrangement of well- characterized poses square
measure contemplated to be utilized in the
correspondence. The different distributions show that
static hand signal acknowledgment remains field of
dynamic examination, while a few of them endeavor to
confront the previously notice disk to improve the
exhibition and nature of existing advances [1].
The use of signal acknowledgment framework
continuously ought to give high precision and heartiness
to the different mess foundations. This paper shows the
advancement of vision-based static hand signal
acknowledgment framework utilizing web camera
continuously applications. The preprocessing stage
comprises of light remuneration, division, filtering, hand
district identification and picture resize. This work
proposes a discrete wavelet change (DWT) and Fisher
proportion (F-proportion) based element extraction
system to characterize the hand motions in an
uncontrolled domain. The exhibition of the proposed
strategy is assessed on two standard open datasets and
one indigenously created complex foundation dataset for
acknowledgment of hand signals [2].
The proposed calculation is autonomous of hand heading
and doesn't utilize any markers or information gloves.
This is generally completed by the technique for division
for static pictures and by the strategy for following for
dynamic pictures. For dynamic signals, the hand motion
should be identified and followed. For hand following,
either the video is separated into outlines and each casing
is prepared alone, or some following subtleties like shape
or skin shading utilizing a few apparatuses [4].
In this paper, new deep learning model is used to
recognize the hand gestures using Convolutional Neural
Network (CNN). The disadvantages is that it focuses only
ISSN 2278
-
3091
Volume 9 No.2, March -April 2020
International Journal of Advanced Trends in Computer Science and Engineering
Available Online at http://www.warse.org/IJATCSE/static/pdf/file/ijatcse106922020.pdf
https://doi.org/10.30534/ijatcse/2020/106922020
C.Selvarathi et al., International Journal of Advanced Trends in Computer Science and Engineering, 9(2), March - April 2020, 1600 – 1603
1601
on static data images. Using another algorithm namely
CTC and FMCW methods we can recognize the dynamic
inputs [7].
This paper shows a computationally efficient technique
for activity acknowledgment from profundity video
arrangements. It utilizes the purported profundity
movement maps (DMMs) from three projection sees
(front, side and top) to catch movement signs and sees
nearby paired examples (LBPs) to increase a minimized
component portrayal. The trial results on two standard
datasets exhibited enhancement over the acknowledgment
exhibitions of the current technique [5].
3. PROPOSEDSYSTEM
This paper shows a computationally efficient technique
for activity acknowledgment from profundity video
arrangements. It utilizes the purported profundity
movement maps (DMMs) from three projection sees to
catch movement signs and uses nearby paired examples
(LBPs) to increase a minimized component portrayal.
The trial results on two standard datasets exhibited
enhancements over the acknowledgment exhibitions of
the current techniques.
3.1 Cascade Classifier
The cascade classifier consists of many stages, where
each stage contains different methods and specific
periods by recognizing the structures of images used.
Each period of the classifier names the specific region in
the window and recognizes both positive and negative
labels. If it shows positive results, it denotes that the
image is recognized and negative shows that images not
found.
4. METHODOLOGY
The methodology of a system uses Region of interest
method that includes identifying regions of the object
appropriately
Figure 1: Methodology for hand recognition
Right now, can utilize either equipment or programming
like camera, sensors for perceiving inputs. Utilizing
foundation and forefront subtraction, we can recognize
the articles. HAAR course calculation perceives the
stationary foundation with the goal that our framework
will have littler quest district for following the
application. By utilizing ROI, tallying the pixels of
pictures and characterize them as per profound learning
strategies.
5. EXPERIMENTAL RESULTS
5.1 Experimental Problem
The problem is gestures are only recognized by certain
amount of people feed in the system.
Figure 2 :Recognition using gloves
5.2 Experimental Results
This system overcomes the above problem by recognizing
the dynamic number of inputs.
Camera
Background
Substraction
Foreground
Substraction
ROI
Finger count
Classify (Deep
Learning)
HAAR cascade
Hand pixels
C.Selvarathi et al., International Journal of Advanced Trends in Computer Science and Engineering, 9(2), March - April 2020, 1600 – 1603
1602
Figure 3: Recognition of gestures
Figure 4:Different types of gestures
Figure 5: ROI Extraction
Figure 6: Finger count detection
6. CONCLUSION
In past examinations, they favor equipment control to
recognize the hands. Hand division become complex of
different foundations Segmentation precision is less close
by following. Presently, introduced a technique to
perceive the obscure information motions by utilizing
hand following and extraction strategy. Apply this
framework to perceive the single signal. In the
investigations, we expect stationary foundation with the
goal that our framework will have littler quest locale for
following the application.
REFERENCES
1 Hua Li, Member IEEE, Lifan Wu, Huan Wang, Cheng
Han, Wei Quan and Jianping Zhao “Hand gesture
recognition enhancement based on spatial fuzzy
matching in Leap Motion” IEEE2019
2 Y. Song, J. Tang, F. Liu, and S. Yan, “Body surface
context: A new robust feature for action recognition
from depth videos,” IEEE Trans-actions on Circuitsan
Systems for Video Technology, vol. 24, no. 6, pp. 952–
964, 2014
https://doi.org/10.1109/TCSVT.2014.2302558
3 M. Ye, Q. Zhang, L. Wang, J. Zhu, R. Yang, and J.
Gall, “A survey on human motion analysis from depth
data, ”Lecture Notes in Computer Science, vol. 8200, pp.
149–187, 2013.
https://doi.org/10.1007/978-3-642-44964-2_8
4.JokoSutopo1,MohdKhanapi,AbdGhani,M.A.Burhanudd
in , Zulhawati ,” Gesture Recognition of Dance using
Chain Code and Hidden Markov Model”,
vol.8,No.6,2019
https://doi.org/10.30534/ijatcse/2019/85862019
5 Sadman Shahriar Alam, AkibJayed Islam, Nahid
Nasrin, KhandokerTanjimAhammad“ Hand Gesture
Detection Using Haar Classifier with Appropriate
Skin Color, Kernal Sizing & Auto Thresholding”
IEEE2017
6 S.Thilagamani , N. Shanthi,” Object Recognition
Based on Image Segmentation and Clustering”,
Journal of Computer Science,Vol. 7,No.11,pp. 1741-
1748, 2011.
https://doi.org/10.3844/jcssp.2011.1741.1748
C.Selvarathi et al., International Journal of Advanced Trends in Computer Science and Engineering, 9(2), March - April 2020, 1600 – 1603
1603
7 R. Y. Wang and J. Popovic, “Real-time hand tracking
with a color glove,” in ACM SIGGRAPH, 2009.
https://doi.org/10.1145/1576246.1531369
8 P.Buehler, M.Everingham, D.Huttenlocher, and A.
Zisserman, “Upper body detection and tracking in
extended signing sequences,” International Journal of
Computer Vision, vol. 95, no. 2, pp. 180–197, 2011.
9 X.Zhang, X.Chen, Y.Li, V.Lantz,K.Wang, and J.
Yang, “A frame-work for hand gesture recognition
based on accelerometer and emg sensors,” IEEE
Transactions on Systems, Man and Cybernetics, Part A:
Systems and Humans, vol. 41, no. 6, pp. 1064–1076,
2011.
10 S. D. Kelly, S. M. Manning, and S. Rodak, “Gesture
gives a hand to language and learning: Perspectives
from cognitive neuroscience, developmental
psychology and education,” Language and Linguistics
Compass, vol. 2, no. 4, pp. 569–588, 2008.
11 N Deepa ,P. Pandiaraja 2019 , Hybrid Context
Aware Recommendation System for E-Health Care by
merkle hash tree from cloud using evolutionary
algorithm , Journal of Soft Computing , Springer 2019.
https://doi.org/10.1007/s00500-019-04322-7
12 P.RajeshKanna and P.Pandiaraja 2019,An Efficient
Sentiment Analysis Approach for Product Review
using Turney Algorithm, Journal of Procedia Computer
Science , Elsevier ,Vol 165 ,Issue 2019, Pages 356-362.
https://doi.org/10.1016/j.procs.2020.01.038
13 P. Santhi, S.Thilagamani, A Survey on Audit Free
Cloud Storage via Deniable Attribute Based
Encryption, IRA-International Journal of Technology &
Engineering,Vol.5,No.1,PP.1-5,2016.
https://doi.org/10.21013/jte.v5.n1.p1
14 S.Thilagamani, N. Shanthi,, Object Recognition
Based on Image Segmentation and Clustering, Journal
of Computer Science, Volume 7,No.11,pp. 1741-1748,
2011.
https://doi.org/10.3844/jcssp.2011.1741.1748
15 P.Santhi,G.Mahalakshmi ,Classification of magnetic
resonance images using eight directions gray level co-
occurrence matrix (8DGLCM) based feature
extraction, International journal of engineering and
advanced technology, volume 8,No.4,pp.839-846,2019.
16 K Sumathi, P Pandiaraja , Dynamic alternate buffer
switching and congestion control in wireless
multimedia sensor networks , Journal of Peer-to-Peer
Networking and Applications , Springer,2019.
https://doi.org/10.1007/s12083-019-00797-1
17 Ahmad al-Qerem,Arwa Alahmad , Human Body
Poses Recognition Using Neural Networks with Data
Augmentation , International Journal of Advanced
Trends in Computer Science and Engineering, Vol
8,Issue 5,September - October 2019, pp 2117 – 2120.
https://doi.org/10.30534/ijatcse/2019/40852019
18 Hisham Mobideen , Sattam Allahawiah,
KhaledMohammad Alomari , The Impact of Human
Resources Information Systems on Human Resources
Selection and Recruitment Strategy: An applied study
on Arab Potash Company in the Hashemite Kingdom
of Jordan, International Journal of Advanced Trends in
Computer Science and Engineering, Vol 8,Issue
5,September - October 2019, pp 2475- 2484.
https://doi.org/10.30534/ijatcse/2019/93852019
... This solution is always based on image vision algorithms. Several algorithms are used for hand segmentation based on a background model [30]. Systems with a color-based algorithm are used to separate the region of an object of interest [5]. ...
... This can generally provide precise results, however the cost of these systems is high. Hardware maintenance and uncomfortable use are also issues in this approach [11,21,30,37] Motion Detection Through frame difference to find out moving hand. [29,35,44] Machine Learning Use Haar-like feature classifier and cluster analysis to judge hand gesture [2,6,7,18,19,31,34] Multimedia Tools and Applications (2020) 79:59 9-6007 [9,22]. ...
Design of hand gesture recognition system for human-computer interaction
Article
Full-text available
  • Mar 2020
  • MULTIMED TOOLS APPL
Human-Computer interaction (HCI) with gesture recognition is designed to recognize a number of meaningful human expressions, and has become a valuable and intuitive computer input technique. Hand gestures are one of the most intuitive and common forms of communication, and can communicate a wide range of meaning. Vision-based hand gesture recognition has received a significant amount of research attention in recent years. However, the field still presents a number of challenges for researchers. In the vision-based hand gesture interaction process between humans and computers, gesture interpretation must be performed quickly and with high accuracy. In this paper, a low-cost HCI system with hand gesture recognition is proposed. This system uses several vision techniques. Skin and motion detection is used for capturing the region-of-interest from the background regions. A connected component labeling algorithm is proposed to identify the centroid of an object. To identify the exact area of hand gesture, the arm area is removed with the aid of a convex hull algorithm. Moreover, a real-time demonstration system is developed, based on a single-camera mechanism which allows for the use of wearable devices. Simulation results show that the recognition rate is still high, although some interference is encountered in the simulated environments.
View
... The first method of collecting data was based on gloves fitted with sensors. The second method was a computer vision approach that used camera vision to allow humans and computers to communicate [Selvarathi et al, 2020]. ...
Interpreting Arabic sign alphabet by utilizing a glove with sensors
Article
Full-text available
  • Aug 2022
People who are deaf or dumb in Arab communities face several challenges. The most important challenge is to communicate with people. In this study, a new approach for identifying the alphabet in the Iraqi Sign Language (IrSL) is proposed, which makes use of a suggested deep neural network called the Deep Recurrent Alphabet Sign Language (DRASL). It utilizes the Long Short-Term Memory (LSTM) technique for classifying the outputs and recognizing the alphabet in the SL. The dataset is constructed with the use of a glove that is coupled to flex sensors on each finger; each sensor gives a variable value based on the curvature ratio of the fingers. The sensors were connected to an Arduino which was then linked to a computer to transfer the data we collected. The data were divided into three groups, which had 29 different movements. All of these groups had a remarkably high accuracy equal to 100%.deep learning
View
... For each point in the real 3D point cloud model, calculate the point with the smallest Euclidean distance from the point in the reconstructed model. e completeness is defined as the number of points in the real 3D model that is less than the given distance threshold, accounting for the total number of points in the real model [8,9]. Since the real results obtained by laser scanning are not necessarily complete, care should be taken to prevent the deviation of the results caused by the actual scene points that may be lost. ...
Human Posture Recognition and Estimation Method Based on 3D Multiview Basketball Sports Dataset
Article
Full-text available
In traditional 3D reconstruction methods, using a single view to predict the 3D structure of an object is a very difficult task. This research mainly discusses human pose recognition and estimation based on 3D multiview basketball sports dataset. The convolutional neural network framework used in this research is VGG11, and the basketball dataset Image Net is used for pretraining. This research uses some modules of the VGG11 network. For different feature fusion methods, different modules of the VGG11 network are used as the feature extraction network. In order to be efficient in computing and processing, the multilayer perceptron in the network model is implemented by a one-dimensional convolutional network. The input is a randomly sampled point set, and after a layer of perceptron, it outputs a feature set of n × 16. Then, the feature set is sent to two network branches, one is to continue to use the perceptron method to generate the feature set of n × 1024, and the other network is used to extract the local features of points. After the RGB basketball sports picture passes through the semantic segmentation network, a picture containing the target object is obtained, and the picture is input to the constructed feature fusion network model. After feature extraction is performed on the RGB image and the depth image, respectively, the RGB feature, the local feature of the point cloud, and the global feature are spliced and fused to form a feature vector of N × 1152. There are three branches for this vector network, which, respectively, predict the object position, rotation, and confidence. Among them, the feature dimensionality reduction is realized by one-dimensional convolution, and the activation function is the ReLU function. After removing the feature mapping module, the accuracy of VC-CNN_v1 dropped by 0.33% and the accuracy of VC-CNN_v2 dropped by 0.55%. It can be seen from the research results that the addition of the feature mapping module improves the recognition effect of the network to a certain extent
View
Virtual mouse with RGB colored tapes
Article
Full-text available
  • Jan 2016
A recent development in growing computer environments has resulted in a push for the development of more intuitive forms of human-computer interaction method [HCI]. The aim of this research is to investigate a new approach for controlling the mouse movements using a real-time camera. Instead of changing or adding mouse parts, a method where access to the mouse/cursor by hand gestures with colored tapes can be enabled. It tries to use a camera and computer vision technologies such as gesture recognition and image segmentation to control mouse tasks with colored tapes and shows how it can perform all the mouse functions a current mouse device can. In the first phase, hand gestures are acquired using a camera based on colour detection technique using segmentation and image subtraction algorithm.In the second phase, RGB colored tapes are used to control different functions of the mouse and also the combination of these 3 colors by considering the area of each object/colored tape using the blob analysis and the bounding box algorithm. The user must wear the red, green and blue tapes to the fingers such that it is easy to make the movements for each tape and also the combination of the colored tapes to acquire the desired output of the cursor movement in the system.
View
Video Switching System Using Non-Contact Interface by Image Processing: Toward an Interactive Digital Signage
Conference Paper
  • Jul 2017
We describe a method to detect human movement using a single camera and construct a non-contact human interface in this paper. In this system, a virtual switch is set in the air, and when the user touches the virtual switch, built-in five videos can be sequentially switched. In order to examine the operation accuracy of the system, a total of 460 evaluation experiments were conducted. The evaluation experiments were conducted on two cases, i.e., the case where the operation was performed without passing through the glass and the case where the operation was performed through the glass. As a result, the success rate in the case of not passing the glass was 95.8%, the success rate in the case of passing the glass was 96.0%, both were a very high recognition success rate. From these result, we can say that this system is sufficiently practical as a human computer interface. Furthermore, it was confirmed that this system can be operated through the glass. This means that this system can be operated from the outside even if it is installed in the show window. When we install the proposed system to a computer with built-in camera, we can easily construct an inexpensive interactive digital signage without any external device.
View
An Efficient Sentiment Analysis Approach for Product Review using Turney Algorithm
Article
Full-text available
  • Jan 2019
Sentiment analysis can be done by means of Classification and its most important tasks are text categorization, tone recognition, image classification etc. Mostly the extant methods of supervised classification are based on traditional statistics, which can provide ideal results. The main aim is to increase the accuracy and to report the manufacturer about the negatives of the product. The major problem is categorization of sentiment polarity, which is the problem of sentiment analysis. There are two levels of categorization and they are Review-level Categorization and Sentence-level Categorization. Categorization of review-level becomes arduous when we attempt to classify the reviews respect with their specific rating related to star-scaled. Second, Review-level Categorization has a drawback in Implicit-level sentiment analysis. Mostly SVM, Naïve Bayesian and Decision Tree are mainly used to improve the efficiency of classification. Amazon Dataset is used as Dataset in proposed system to improve the accuracy of Turney algorithm. Semantic Orientation (SO) with Point wise Mutual Information yields good results than other classification methods. The review level gets subjected as positive value, on acquaintance of positive average SO. On the other hand, the review level acquires a negative level in accordance with attainment of negative average SO.
View
View
Hybrid Context Aware Recommendation System for E-Health Care by merkle hash tree from cloud using evolutionary algorithm
Article
Full-text available
  • May 2020
  • SOFT COMPUT
Privacy preservation permits doctors to outsource the huge encrypted reports to the cloud and permits the authenticated patients to have a safe search over the reports without leaking the private information. The doctors in our proposed have used the merkle hash tree for storing the reports of all the patients in the hospital. The existing schemes have used many types of trees like binary tree, red–black tree, spanning tree, B⁺ tree, etc., for the index generation purpose. Since the security is less and the searching time is high for the above said trees, we have proposed the index generation phase based on the merkle hash tree based on the evolutionary algorithm and it takes less time for searching and highly secure for storing the patient reports. The evolutionary algorithm is used for breeding the new data’s through crossover as well as mutation operations to give confinement to new children. When the patient submits the search request for specialized doctor, based on the patient disease our protocol will recommend the specialized doctors and send the recommended doctors information to the patients who have the highest rating in the online social networks. After receiving the recommended results, the patient can have the treatment via online booking appointment, video call or in person based on the appointment booked. After completely cured, the patients can rate the doctors based on the medicine satisfaction, doctors’ fees and doctor’s response over the call. In this mechanism, we have used the hybrid context aware recommendation system collaborative filtering for rating the doctors based on their performance. After rating the doctors, our protocol has measured the accuracy based on the predicted rating and the true rating. This kind of accuracy metrics is used for ranking the good doctors in the top rank for the patient use. Our proposed work Hybrid Context Aware Recommendation System for E-Health Care (HCARS-EHC) is implemented, and the implementation results of HCARS-EHC illustrate that our protocol is efficient based on the privacy preservation, recommendation and ranking with less computation and communication complexity.
View
Dynamic alternate buffer switching and congestion control in wireless multimedia sensor networks
Article
Full-text available
  • Nov 2020
Normally sensor nodes are located at unattended environment which has low battery power. The wireless hypermedia sensor based networks have huge capacity and continuous flow of data from source to sink. Congestion arises at the time of incoming data is greater than available resources. The congestion causes packet loss, buffer over flow, energy wastage and increases end to end delay. The proposed method Dynamic Alternate Buffer switching and Congestion Control is used to manage congestion in an efficient manner. In which congestion is detected based on remaining buffer, remaining energy and trust level of the sensor nodes. This approach is based on cost evaluation method which decides primary and spare buffers at run time. This method is used to enable dynamic buffer switching and optimize the effect of congestion. The performance of BETCC compared with various existing protocol like TCEER and TFCC in terms of energy consumption and data loss ratio.
View
Hand Gesture Detection Using Haar Classifier with Appropriate Skin Color, Kernal Sizing & Auto Thresholding
Article
Full-text available
  • Mar 2017
Hand gesture detection & recognition intends to detect & recognize some meaningful hand sign. It is an utmost challenge to design such intelligent human-computer interface. 3D hand gesture recognition, one of the most advanced technologies for building smart communication method with computers, has been developing an enormous research interest in computer vision, pattern recognition and human-computer interaction (HCI) system. The developing depth sensors enormously propelled different hand motion identification methodologies and applications, which were extremely restricted in the 2D area with conventional cameras. In this paper, we provided a productive method for image preprocessing and detection of certain gesture, advance classifier theory based on genetic algorithm, boosting algorithm & Fischer's linear discriminant algorithm. Improvising the traditional preprocessing of image for gesture detection where skin color detection, kernel matrix evolution & auto-thresholding of image are the most challenging part of this thesis paper. This paper additionally displays a review of the state-of-the-art research for 3D hand motion recognition in four perspectives: 3D hand modeling, basic sensors capable for detecting hand gesture, static hand motion acknowledgment, hand direction motion acknowledgment, continuous hand gesture recognition and related applications of hand gesture.
View
A Survey on Audit Free Cloud Storage via Deniable Attribute Based Encryption
Article
Full-text available
  • Nov 2016
div> Cloud computing is a rising technology which provider an assortment of opportunities for online distribution of resources or services. The most effective advantage of using cloud computing is higher availability of services with less cost and simple scalability. While the storage space of shared data on remote servers is not a new development, current development of cloud computing validates a more careful look at its actual consequences involving privacy and confidentiality issues. As users no longer actually possess the storage of their data, traditional cryptographic primitives for the purpose of data protection cannot be directly accepted. In particular, simply downloading all the data for its integrity confirmation is not a realistic explanation due to the expensiveness in I/O and transmission cost across the network. Besides, it is often not enough to detect the data corruption only when contacting the data, as it does not offer users correctness assurance for those un-accessed information and might be too late to recover the data loss or damage. To fully make sure the data integrity and save the cloud users’ calculation resources as well as online burden, it is of critical importance to allow public auditing service for cloud data, so that users may choice to an independent third party auditor (TPA) to audit the contract out data when needed. The TPA, who has expertise and abilities that users do not, can occasionally check the honesty of all the data stored in the cloud on behalf of the users, which provides a much more better and reasonable way for the users to ensure their storage rightness in the cloud. In a word, allowing public auditing services will play an important role for this emerging cloud market to become fully recognized; where users will need ways to evaluate risk and gain hope in the cloud. </div
View
Object Recognition Based on Image Segmentation and Clustering
Article
Full-text available
  • Nov 2011
  • J Comput Sci
Problem statement: This study deals with object recognition based on image segmentation and clustering. Acquiring prior information of an image is done via two separate processes. Approach: The first process deals with detecting object parts of an image and integration of detected parts into several clusters. All these cluster centers form the visual words. The second process deals with over segmenting the image into super pixels and formation of larger sub region using Mid-level clustering algorithm, since it incorporates various information to decide the homogeneity of a sub region. Results: The outcome of the two processes are used for the similarity graph representation for object segmentation as proposed. In order to model the relationship between the shape and color or texture matrix representation has been used. Mask map ensures that the probability of each super pixel to harp inside an object. Conclusion: The basic whim is to integrate all the priors into an uniform framework. Thus the ORBISC can handle size, color, texture and pose variations better than those methods that focus on the objects only.
View
Human Body Poses Recognition Using Neural Networks with Data Augmentation
Article
  • Oct 2019
The continuous development of technology of IT enabled computers to see and learn. There are many viable applications for computer learning and vision to solve new tasks. In this paper, we propose a framework, able of automatically perceiving the human body poses from a single image, acquired by a traditional low-cost camera. Our methodology exploiting the vision of computers features and neural networks to detect and recognize a human from an image. The processing start with detection human in image and then extracting the silhouette from an image then using a neural network to recognize body poses based on silhouettes that extracted. To classified the detected silhouettes with body poses, the neural network was trained with dataset of preprocessed images contains silhouettes and labelled using 0, 1 for standing and walking poses respectively labelled , according to our outcomes that resulted in an accuracy over 93% with the best neural network model. The proposed approach offers promising effects with accepted accuracy. This knowledge will help greatly in decision-making and provide independent vehicles with valuable information for their artificial intelligence to process and make decisions as regards the movement of pedestrians.
View
Hand Gesture Recognition Enhancement Based on Spatial Fuzzy Matching in Leap Motion
Article
  • Jul 2019
Gesture recognition is an important human- computer interaction interface. This paper introduces a novel hand gesture recognition system based on Leap Motion gen. 2. In this system, a spatial fuzzy matching (SFM) algorithm is firstly presented by matching and fusing spatial information to construct a fused gesture dataset. For dynamic hand recognition, an initial frame correction strategy based on SFM is proposed to fast initialize the trajectory of test gesture with respect to the gesture dataset. A notable feature of this system is that it can run on ordinary laptops due to the small size of the fused dataset, which accelerates the calculation of recognition rate. Experiment results show that the system recognizes static hand gestures at recognition rates of 94%-100% and over 90% of dynamic gestures using our collected dataset. This can greatly enhance the usability of Leap Motion.
View
A Survey on Human Motion Analysis from Depth Data
Chapter
  • Jan 2013
Human pose estimation has been actively studied for decades. While traditional approaches rely on 2d data like images or videos, the development of Time-of-Flight cameras and other depth sensors created new opportunities to advance the field. We give an overview of recent approaches that perform human motion analysis which includes depth-based and skeleton-based activity recognition, head pose estimation, facial feature detection, facial performance capture, hand pose estimation and hand gesture recognition. While the focus is on approaches using depth data, we also discuss traditional image based methods to provide a broad overview of recent developments in these areas.
View