Chongqing Technical and Business Institute

Chongqing, China

Chongqing Technical and Business Institute

Chongqing, China
SEARCH FILTERS
Time filter
Source Type

Geng Z.,Chongqing Technical and Business Institute | Hao X.,Chongqing Technical and Business Institute | Wu W.,Chongqing Technical and Business Institute
Journal of Convergence Information Technology | Year: 2012

This paper proposed a image feature extraction and recognition method using various image recognition techniques such as the binary image boundary point searching algorithm, the determination of face exterior rectangle based on skin color segmentation and geometric feature, the eye location algorithm based on chrome space and mouth location algorithm using SUSAN corner detection method. This method can reduce computation region as well as computational complexity and other disturbance factors. Therefore, it has established good foundation for next application. The experimental results show that this method has achieved high performance in speed, efficiency and accuracy.


Geng Z.,Chongqing Technical and Business Institute | Hao X.,Chongqing Technical and Business Institute | Wu W.,Chongqing Technical and Business Institute
International Journal of Digital Content Technology and its Applications | Year: 2012

Vocal print recognition is a technology that recognition the speaker's identity by taking advantage of voice parameter reflecting the physiological and behavior characteristic of the voice wave. The primary voice parameters are MFCC LPCC and so on. The classical model is the GMM model. This text bring forward the view point that replacing the GMM by the GRNNs that constructed by the MFCC coefficients and the keynote pitch of the sonant frame data. After analysis of the recognition process and do some experiment, the conclusion is that the confidence sore of mode basing on the GRNNs is much better tan the GMM mode.


Geng Z.,Chongqing Technical and Business Institute | Hao X.,Chongqing Technical and Business Institute | Kefeng S.,Chongqing Technical and Business Institute
International Journal of Digital Content Technology and its Applications | Year: 2012

In this paper, we proposed and implemented Z-MHTR, a multipath extension of the ZigBee hierarchical tree routing. We presented a preliminary performance study of the proposed node disjoint multipath routing for cluster-tree ZigBee WSNs via practical experiments. Results showed that, generally, under higher data rates, the multipath routing performance degrades. Because, the large number of collisions at the sink and the intra-path collisions due to the hidden node problem not dealt by ZigBee and the intra-path and inter-path interference all contribute to a further decrease in the available throughput and an increase of the data time transmission in ZigBee networks.

Loading Chongqing Technical and Business Institute collaborators
Loading Chongqing Technical and Business Institute collaborators