Image texture classification using wavelet based curve fitting and probabilistic neural network

This work reports on convolutional neural network (CNN) performance on an image texture classification task as a function of linear image processing and number of training images. Detection performance of single and multi-layer CNNs (sCNN/mCNN) are compared to optimal observers. Performance is quantified by the area under the receiver operating characteristic (ROC) curve, also known as the AUC. For perfect detection AUC = 1.0 and AUC = 0.5 for guessing. The Ideal Observer (IO) maximizes AUC but is prohibitive in practice because it depends on high-dimensional image likelihoods. The IO performance is invariant to any fullrank, invertible linear image processing. This work demonstrates the existence of full-rank, invertible linear transforms that can degrade both sCNN and mCNN even in the limit of large quantities of training data. A subsequent invertible linear transform changes the images’ correlation structure again and can improve this AUC. Stationary textures sampled from zero mean and unequal covariance Gaussian distributions allow closed-form analytic expressions for the IO and optimal linear compression. Linear compression is a mitigation technique for high-dimension low sample size (HDLSS) applications. By definition, compression strictly decreases or maintains IO detection performance. For small quantities of training data, linear image compression prior to the sCNN architecture can increase AUC from 0.56 to 0.93. Results indicate an optimal compression ratio for CNN based on task difficulty, compression method, and number of training images.

Download Full-text

Classification and Identification of Power System Disturbances Using Wavalet and Artificial Neural Network Technique

Scientific Research Journal ◽

10.24191/srj.v2i2.9330 ◽

2005 ◽

Vol 2 (2) ◽

pp. 25

Author(s):

Noraliza Hamzah ◽

Wan Nor Ainin Wan Abdullah ◽

Pauziah Mohd Arsad

Keyword(s):

Neural Network ◽

Power Quality ◽

Probabilistic Neural Network ◽

Microsoft Excel ◽

Waveform Data ◽

The Neural Network ◽

Harmonic Voltage ◽

Power Quality Disturbances ◽

Electronic Load ◽

Wavelet Technique

Power Quality disturbances problems have gained widespread interest worldwide due to the proliferation of power electronic load such as adjustable speed drives, computer, industrial drives, communication and medical equipments. This paper presents a technique based on wavelet and probabilistic neural network to detect and classify power quality disturbances, which are harmonic, voltage sag, swell and oscillatory transient. The power quality disturbances are obtained from the waveform data collected from premises, which include the UiTM Sarawak, Faculty of Science Computer in Shah Alam, Jati College, Menara UiTM, PP Seksyen 18 and Putra LRT. Reliable Power Meter is used for data monitoring and the data is further processed using the Microsoft Excel software. From the processed data, power quality disturbances are detected using the wavelet technique. After the disturbances being detected, it is then classified using the Probabilistic Neural Network. Sixty data has been chosen for the training of the Probabilistic Neural Network and ten data has been used for the testing of the neural network. The results are further interfaced using matlab script code. Results from the research have been very promising which proved that the wavelet technique and Probabilistic Neural Network is capable to be used for power quality disturbances detection and classification.

Download Full-text