A Novel Illumination Compensation Method Using Wavelet Packet Transformation

2011 ◽  
Vol 317-319 ◽  
pp. 897-900
Author(s):  
Zhen Hua Zhao ◽  
Xiao Hong Hao
Keyword(s):  
Multiplicative Noise ◽  
Additive Noise ◽  
Recognition Performance ◽  
Wavelet Packet ◽  
Recognition Rate ◽  
Face Image ◽  
Experimental Results ◽  
Illumination Compensation ◽  
Wavelet Packet Transformation ◽  
Facial Images

A novel illumination compensate method is proposed in this paper to improve recognition performance. A modified lighting model called Lambertin which includes additive noise and multiplicative noise are presented firstly. Then, additive noise is removed by using wavelet packet transformation. Next, the processed image is transformed into logarithm domain and the multiplicative noise, which has been named additive noise, is removed by means of the same above algorithm. Finally, a compensated face image is obtained. We examine the proposed method on Yale extended B database compared with other methods. Experimental results show that our algorithm improves by 3%~12% recognition rate. It can effectively adjust the facial images for varying illumination conditions and also improve the recognition performance and robustness.

A LOW-COMPUTATION APPROACH FOR HUMAN FACE RECOGNITION

2012 ◽  
Vol 26 (06) ◽  
pp. 1256015 ◽  
Author(s):  
WEI-LI FANG ◽  
YING-KUEI YANG ◽  
JUNG-KUEI PAN
Keyword(s):  
Feature Extraction ◽  
Face Recognition ◽  
Recognition Performance ◽  
Recognition Rate ◽  
Face Image ◽  
Experimental Results ◽  
Human Face ◽  
Human Face Recognition ◽  
Computation Cost ◽  
Recognition Algorithms

Several 2DPCA-based face recognition algorithms have been proposed hoping to achieve the goal of improving recognition rate while mostly at the expense of computation cost. In this paper, an approach named SI2DPCA is proposed to not only reduce the computation cost but also increase recognition performance at the same time. The approach divides a whole face image into smaller sub-images to increase the weight of features for better feature extraction. Meanwhile, the computation cost that mainly comes from the heavy and complicated operations against matrices is reduced due to the smaller size of sub-images. The reduced amount of computation has been analyzed and the integrity of sub-images has been discussed thoroughly in the paper. The experiments have been conducted to make comparisons among several better-known approaches and SI2DPCA. The experimental results have demonstrated that the proposed approach works well on reaching the goals of reducing computation cost and improving recognition performance simultaneously.

scholarly journals DoPAMINE: Double-Sided Masked CNN for Pixel Adaptive Multiplicative Noise Despeckling

2019 ◽  
Vol 33 ◽  
pp. 4031-4038
Author(s):  
Sunghwan Joo ◽  
Sungmin Cha ◽  
Taesup Moon
Keyword(s):  
Neural Network ◽  
Multiplicative Noise ◽  
Additive Noise ◽  
State Of The Art ◽  
Unbiased Estimate ◽  
Experimental Results ◽  
Fine Tuning ◽  
Noisy Image ◽  
Supervised Training ◽  
The Given

We propose DoPAMINE, a new neural network based multiplicative noise despeckling algorithm. Our algorithm is inspired by Neural AIDE (N-AIDE), which is a recently proposed neural adaptive image denoiser. While the original NAIDE was designed for the additive noise case, we show that the same framework, i.e., adaptively learning a network for pixel-wise affine denoisers by minimizing an unbiased estimate of MSE, can be applied to the multiplicative noise case as well. Moreover, we derive a double-sided masked CNN architecture which can control the variance of the activation values in each layer and converge fast to high denoising performance during supervised training. In the experimental results, we show our DoPAMINE possesses high adaptivity via fine-tuning the network parameters based on the given noisy image and achieves significantly better despeckling results compared to SAR-DRN, a state-of-the-art CNN-based algorithm.

STUDY OF PERCEPTUAL SIMILARITY BETWEEN DIFFERENT LEXICONS

2004 ◽  
Vol 18 (07) ◽  
pp. 1321-1338 ◽  
Author(s):  
CINTHIA O. A. FREITAS ◽  
FLÁVIO BORTOLOZZI ◽  
ROBERT SABOURIN
Keyword(s):  
Visual Perception ◽  
Error Rate ◽  
Recognition Performance ◽  
Recognition Rate ◽  
Experimental Results ◽  
Perceptual Similarity ◽  
Training Set ◽  
Feature Similarity ◽  
Observation Sequence ◽  
French Words

The study investigates the perceptual feature similarity between different lexicons based on visual perception of the words and their representation through an observation sequence. We confirm that it is possible to use databases, which are similar in terms of morphological/perceptual features to improve the recognition performance. In this work, we demonstrated through experimentation, that it is possible to improve the recognition rate of handwritten Portuguese words by adding samples of French words in the training set. Experimental results show the efficiency of this strategy reducing the error rate.

An Illumination Compensation Algorithm Based on Single-Scale Retinex and its Application in Face Recognition

2012 ◽  
Vol 546-547 ◽  
pp. 755-759
Author(s):  
Hong Hai Liu ◽  
Xiang Hua Hou
Keyword(s):  
Face Recognition ◽  
Recognition Rate ◽  
Gaussian Function ◽  
Vertical Direction ◽  
Face Image ◽  
Enhancement Effect ◽  
Original Image ◽  
Illumination Compensation ◽  
Compensation Algorithm ◽  
Reflection Image

Due to the affect of illumination conditions when taking face image, the traditional algorithm effect of face recognition in practice is unsatisfied. Thus, in order to reduce the affect of illumination conditions, the pretreatment of illumination compensation to face image is needed. The common illumination compensation algorithms have two kinds: linear compensation and nonlinear compensation, such as histogram equalization method, log algorithm, and so on, but those algorithms only perform localized enhancement for image and they cannot really reflect the original image. In this paper, a kind of illumination compensation algorithm based on color constant theory is put forward. An image is mainly made of reflection image and incident image convolution. With color constancy, the incident image is not subject to the influence of illumination conditions and reflection image weakens image effect. Thus, if the incident image is found, the image is enhanced. Firstly, the acquisition method of incident image is analyzed. By the decomposition of Gaussian function, the incident image can be obtained, which provides a theoretical basis for the obtainment of incident image. Then, the expression of Retinex is analyzed. Secondly, we analyze the algorithm based on Retinex. In the horizontal direction and vertical direction of image, compensation is performed and the image after compensation is expanded, so that the compensation image is obtained. Lastly, the illumination compensation results of face image by face image database CMU are given. Compared with the original image, the results show that there is obvious enhancement effect. The recognition rate of face image in complex illumination conditions is improved.

scholarly journals Recognizing Human Races through Machine Learning—A Multi-Network, Multi-Features Study

Mathematics ◽  
2021 ◽  
Vol 9 (2) ◽  
pp. 195
Author(s):  
Adrian Sergiu Darabant ◽  
Diana Borza ◽  
Radu Danescu
Keyword(s):  
Race And Ethnicity ◽  
Large Scale ◽  
Recognition Rate ◽  
Social Signals ◽  
Image Characteristics ◽  
Race Effect ◽  
Race Ethnicity ◽  
The Impact ◽  
Facial Images ◽  
Ethnicity Classification

The human face holds a privileged position in multi-disciplinary research as it conveys much information—demographical attributes (age, race, gender, ethnicity), social signals, emotion expression, and so forth. Studies have shown that due to the distribution of ethnicity/race in training datasets, biometric algorithms suffer from “cross race effect”—their performance is better on subjects closer to the “country of origin” of the algorithm. The contributions of this paper are two-fold: (a) first, we gathered, annotated and made public a large-scale database of (over 175,000) facial images by automatically crawling the Internet for celebrities’ images belonging to various ethnicity/races, and (b) we trained and compared four state of the art convolutional neural networks on the problem of race and ethnicity classification. To the best of our knowledge, this is the largest, data-balanced, publicly-available face database annotated with race and ethnicity information. We also studied the impact of various face traits and image characteristics on the race/ethnicity deep learning classification methods and compared the obtained results with the ones extracted from psychological studies and anthropomorphic studies. Extensive tests were performed in order to determine the facial features to which the networks are sensitive to. These tests and a recognition rate of 96.64% on the problem of human race classification demonstrate the effectiveness of the proposed solution.

Online Biological Literature MRI Image Classification Study

2013 ◽  
Vol 694-697 ◽  
pp. 2336-2340
Author(s):  
Yun Feng Yang ◽  
Feng Xian Tang
Keyword(s):  
Recognition Performance ◽  
Recognition Rate ◽  
Image Features ◽  
Feature Identification ◽  
Identification Method ◽  
Image Library ◽  
Text Information ◽  
Correct Recognition Rate ◽  
Biological Literature ◽  
Mri Image

In order to construct a certain standard structure MRI (Magnetic resonance imaging) image library by extracting and collating unstructured literature data information, an identification method of the image and text information fusion is proposed. The method makes use of PHOW (Pyramid Histogram Of Words) to represent image features, combines with the word frequency characteristics of the embedded icon note (text), and then uses posterior multiplication fusion method to complete the classification and identification of the online biological literature MRI image. The experimental results show that this method has better correct recognition rate and better recognition performance than feature identification method only based on PHOW or text. The study can offer use for reference to construct other structured professional database from online literature.

Palm Vein Recognition Based on Independent Component Analysis

2013 ◽  
Vol 333-335 ◽  
pp. 1106-1109
Author(s):  
Wei Wu
Keyword(s):  
Pattern Recognition ◽  
Independent Component Analysis ◽  
Euclidean Distance ◽  
Recognition Performance ◽  
Component Analysis ◽  
Independent Component ◽  
Experimental Results ◽  
Vein Pattern ◽  
Vein Recognition ◽  
Palm Vein

Palm vein pattern recognition is one of the newest biometric techniques researched today. This paper proposes project the palm vein image matrix based on independent component analysis directly, then calculates the Euclidean distance of the projection matrix, seeks the nearest distance for classification. The experiment has been done in a self-build palm vein database. Experimental results show that the algorithm of independent component analysis is suitable for palm vein recognition and the recognition performance is practical.

scholarly journals GPS Signal Detection under Multiplicative and Additive Noise

2012 ◽  
Vol 66 (4) ◽  
pp. 479-500 ◽  
Author(s):  
P. Huang ◽  
Y. Pi ◽  
I. Progri
Keyword(s):  
Signal Detection ◽  
Urban Areas ◽  
Multiplicative Noise ◽  
Additive Noise ◽  
Power Level ◽  
Signal Propagation ◽  
Signal Acquisition ◽  
Detection Scheme ◽  
Additive And Multiplicative Noise ◽  
Gps Signal Acquisition

In some Global Positioning System (GPS) signal propagation environments, especially in the ionosphere and urban areas with heavy multipath, GPS signal encounters not only additive noise but also multiplicative noise. In this paper we compare and contrast the conventional GPS signal acquisition method which focuses on handling GPS signal acquisition with additive noise, with the enhanced GPS signal processing under multiplicative noise by proposing an extension of the GPS detection mechanism, to include the GPS detection model that explains detection of the GPS signal under additive and multiplicative noise. For this purpose, a novel GPS signal detection scheme based on high order cyclostationarity is proposed. The principle is introduced, the GPS signal detection structure is described, the ambiguity of initial PseudoRandom Noise (PRN) code phase and Doppler shift of GPS signal is analysed. From the simulation results, the received GPS signal at low power level, which is degraded by additive and multiplicative noise, can be detected under the condition that the received block of GPS data length is at least 1·6 ms and sampling frequency is at least 5 MHz.

Sign in / Sign up

Export Citation Format

Share Document