http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Face Recognition by Classification using Radial Basis Function
Venkata Naresh Mandhala,Debnath Bhattacharyya,Tai-hoon Kim 보안공학연구지원센터 2015 International Journal of Multimedia and Ubiquitous Vol.10 No.5
The face recognition task involves extraction of unique features from the human face. Manifold learning methods are proposed to project the original data into a lower dimensional feature space by preserving the local neighborhood structure. PCA and LDA are used for the feature extraction and the resultant feature vectors are fused with the different fusion techniques and the proposed method yields good recognition rate with PCA Fusion of PCA and LDA features and those are classified with neural network. In general the size of the face database is too high and it needs more memory and needs more time for training so that to improve time and space complexities there is a need for dimensionality reduction. The extracted features are classified with Neural Network to improve the recognition rate.
Face Detection using Image Morphology – A Review
Venkata Naresh Mandhala,Debnath Bhattacharyya,Tai-hoon Kim 보안공학연구지원센터 2016 International Journal of Security and Its Applicat Vol.10 No.4
This paper presents an approach by various Algorithms stage by stage for Face Detection. It first detects the face portion, removing all other portion from an image. First it will remove the background and then the body or cloths portion of the image. To achieve this we propose an algorithm based on K-Mean clustering, Bresenham’s algorithm, Graham Scan Algorithm. With the help of image morphology, algorithm will detect the skin texture of the face. Image opening and closing will help to get the skin texture. Using a nose template algorithm will find the nose tip by template matching method. Feature vectors are calculated with respect to the nose tip as the origin of 8 octants. Image moment provides a measure for variation in the skin. This total process is done by the feature extraction algorithm.
Jackknife Estimation of Incomplete Data for Data Marts for Customer Relationship Management
Venkata Naresh Mandhala,Lakshmipathi Anantha,Vijay Krishna Dhulipalla,Hye-Jin Kim 보안공학연구지원센터 2016 International Journal of u- and e- Service, Scienc Vol.9 No.9
The commitment of measurements to information mining can be followed back to the work by Bayes in 1763. The business organizations gather information and offer it to the Data Marts. The individuals who run little and medium association needs to set up information warehousing to touch base, best case scenario arrangement. Such datasets contain part of missing qualities, at some point the missing qualities range from 10% to 33%. A portion of the information might be fundamental; to recall such information is a troublesome undertaking and this kind of datasets won't yield better arrangement, to take care of this issue the Expectation Maximization (EM) calculation gauges missing qualities. Utilizing EM Algorithm the outcomes are supplanted in the missing positions of the specific information which serves to exact conclusion. In this paper, point estimators were connected, among which EM calculation gives best gauge. It is watched that the more straightforward models by and large yield the best results.
An Association Rule hiding Algorithm for Privacy Preserving Data Mining
K. Srinivasa Rao,Venkata Naresh Mandhala,Debnath Bhattacharyya,Tai-hoon Kim 보안공학연구지원센터 2014 International Journal of Control and Automation Vol.7 No.10
Privacy preserving data mining is a research area concerned with the privacy driven from personally identifiable information when considered for data mining. This paper addresses the privacy problem by considering the privacy and algorithmic requirements simultaneously. The objective of this paper is to implement an association rule hiding algorithm for privacy preserving data mining which would be efficient in providing confidentiality and improve the performance at the time when the database stores and retrieves huge amount of data. This paper compares the performance of proposed algorithm with the two existing algorithms namely ISL and DSR.
Shade Interest Points for Dynamic Stream Object Categorization
S. Suresh Babu,Venkata Naresh Mandhala,Siva Koteswara Rao Chinnam,Tai-hoon Kim 보안공학연구지원센터 2015 International Journal of Multimedia and Ubiquitous Vol.10 No.4
Discovery of investment focuses for ensuing handling is one of the fundamental parts of machine vision. Object order of pictures vigorously depends on investment point identification from which nearby picture descriptors are registered for picture matching. Since investment focuses are focused around luminance, past methodologies generally overlooked the color viewpoint. Later an approach that uses saliency-based peculiarity determination improved by a primary part dissection based scale choice strategy is created. It is utilized to lessen the affectability to changing imaging conditions, and hence it is a light-invariant investment point's location framework. Utilization of color expands the uniqueness of investment focuses. In the setting of item distinguishment, the human observation framework is regularly pulled in by contrasts between parts of pictures and by movement or moving articles. In this manner, in the feature indexing system, investment focuses give more helpful data when contrasted with static pictures. So we propose to amplify the above methodology for element feature streams utilizing Space-Time Interest Points (Stips) that uses a calculation for scale adaption of spatio-worldly investment focuses. STIP distinguishes moving questions in features and describes some particular changes in the development of these articles. A handy execution of the proposed framework accepts our case to help element streams and further it could be utilized as a part of uses, for example, Motion Tracking, Entity Detection and Naming applications.
Automatic Instrumental Raaga – A Minute Observation to Find Out Discrete System for Carnatic Music
B. Tarakeswara Rao,Venkata Naresh Mandhala,Debnath Bhattacharyya,Tai-hoon Kim 보안공학연구지원센터 2015 International Journal of Multimedia and Ubiquitous Vol.10 No.6
The objective of this paper is to evolve a system, which automatically mines the raaga of an Indian Classical Music. In the first step Note transcription is applied on a given audio file in order to generate the sequence of notes which are used to play the song. In the next step, the features related to Arohana – Avarohana are extracted. The features of two/three songs are then selected in random and given as input to the training system. Totally songs of 72 melakartha raagas and 45 janya raagas are considered. Subsequently, work testing is done by extracting features of one or two songs of each raaga, which are given as inputs in the training part. The generated output indicates the identification of each raaga. Unique labeling has been done for each raaga, for the system to identify the set of trained raagas. In this work 7 instruments namely Veena, Saxophone, Violin, Nadaswaram, Mandolin, Flute and Piano are used. The database generated is trained and tested by using (1) Gaussian Mixed Model (2) Hidden Markov Model (3) K-Nearest Neighbor using Cosine distance and Earth Mover Distance to draw appropriate conclusions.