The transmission of video content over a network raises various issues relating to copyright authenticity,ethics,legality,and privacy.The protection of copyrighted video content is a significant issue in the video ind...The transmission of video content over a network raises various issues relating to copyright authenticity,ethics,legality,and privacy.The protection of copyrighted video content is a significant issue in the video industry,and it is essential to find effective solutions to prevent tampering and modification of digital video content during its transmission through digital media.However,there are stillmany unresolved challenges.This paper aims to address those challenges by proposing a new technique for detectingmoving objects in digital videos,which can help prove the credibility of video content by detecting any fake objects inserted by hackers.The proposed technique involves using two methods,the H.264 and the extraction color features methods,to embed and extract watermarks in video frames.The study tested the performance of the system against various attacks and found it to be robust.The evaluation was done using different metrics such as Peak-Signal-to-Noise Ratio(PSNR),Mean Squared Error(MSE),Structural Similarity Index Measure(SSIM),Bit Correction Ratio(BCR),and Normalized Correlation.The accuracy of identifying moving objects was high,ranging from 96.3%to 98.7%.The system was also able to embed a fragile watermark with a success rate of over 93.65%and had an average capacity of hiding of 78.67.The reconstructed video frames had high quality with a PSNR of at least 65.45 dB and SSIMof over 0.97,making them imperceptible to the human eye.The system also had an acceptable average time difference(T=1.227/s)compared with other state-of-the-art methods.展开更多
In order to reduce redundant empty bin capacity in the probability representation,we present a new color feature arrangement mechanism for mean shift tracking objects.In the proposed mechanism,the important optimal co...In order to reduce redundant empty bin capacity in the probability representation,we present a new color feature arrangement mechanism for mean shift tracking objects.In the proposed mechanism,the important optimal color,or we call it optimal color vector,is clustered by closing Euclidean distance which happens inside the original RGB color 3-D spatial domain.After obtaining clustering colors from the reference image RGB spatial domain,novel clustering groups substitute for original color data.So the new color substitution distribution is as similar as the original one.And then target region in the candidate frame is mapped by the constructed optimal clustering colors and the cluster Indices.In the final,mean shift algorithm gives a performance in the new optimal color distribution.Comparison under the same circumstance between the proposed algorithm and conventional mean shift algorithm shows that the former has a certain advantage in computation cost.展开更多
In the process of tracking the target of the pig,with the change of the size of the tracking target in the video image,the estimated tracking target scale cannot be adaptively updated in real-time,resulting in the low...In the process of tracking the target of the pig,with the change of the size of the tracking target in the video image,the estimated tracking target scale cannot be adaptively updated in real-time,resulting in the low accuracy of the tracking target.In this study,a multi-channel color feature adaptive fusion algorithm was proposed,and the target scale of the pig was updated in real-time by utilizing the contour information of the target pig.Experiments show that the proposed algorithm had a distance precision of 89.7%and an overlap precision of 87.5%,and the average running speed of this algorithm was 50.1 fps.The robustness of the proposed algorithm in tracking target deformation and scale variation were significantly improved,which satisfies the accuracy and real-time requirements of pig target tracking.展开更多
Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the cl...Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the clock.The driver’s face images were captured by a camera with a colored lens and an infrared lens mounted above the dashboard.The landmarks of the driver’s face were labeled and the eye-area was segmented.By calculating the aspect ratios of the eyes,the duration of eye closure,frequency of blinks and PERCLOS of both colored and infrared,fatigue can be detected.Based on the change of light intensity detected by a photosensitive device,the weight matrix of the colored features and the infrared features was adjusted adaptively to reduce the impact of lighting on fatigue detection.Video samples of the driver’s face were recorded in the test vehicle.After training the classification model,the results showed that our method has high accuracy on driver fatigue detection in both daytime and nighttime.展开更多
Image matching based on scale invariant feature transform(SIFT) is one of the most popular image matching algorithms, which exhibits high robustness and accuracy. Grayscale images rather than color images are genera...Image matching based on scale invariant feature transform(SIFT) is one of the most popular image matching algorithms, which exhibits high robustness and accuracy. Grayscale images rather than color images are generally used to get SIFT descriptors in order to reduce the complexity. The regions which have a similar grayscale level but different hues tend to produce wrong matching results in this case. Therefore, the loss of color information may result in decreasing of matching ratio. An image matching algorithm based on SIFT is proposed, which adds a color offset and an exposure offset when converting color images to grayscale images in order to enhance the matching ratio. Experimental results show that the proposed algorithm can effectively differentiate the regions with different colors but the similar grayscale level, and increase the matching ratio of image matching based on SIFT. Furthermore, it does not introduce much complexity than the traditional SIFT.展开更多
The generic Meanshift is susceptible to interference of background pixels with the target pixels in the kernel of the reference model, which compromises the tracking performance. In this paper, we enhance the target c...The generic Meanshift is susceptible to interference of background pixels with the target pixels in the kernel of the reference model, which compromises the tracking performance. In this paper, we enhance the target color feature by attenuating the background color within the kernel through enlarging the pixel weightings which map to the pixels on the target. This way, the background pixel interference is largely suppressed in the color histogram in the course of constructing the target reference model. In addition, the proposed method also reduces the number of Meanshift iterations, which speeds up the algorithmic convergence. The two tests validate the proposed approach with improved tracking robustness on real-world video sequences.展开更多
Automated grading of colon biopsy images across all magnifications is challenging because of tailored segmentation and dependent features on each magnification.This work presents a novel approach of robust magnificati...Automated grading of colon biopsy images across all magnifications is challenging because of tailored segmentation and dependent features on each magnification.This work presents a novel approach of robust magnification-independent colon cancer grading framework to distinguish colon biopsy images into four classes:normal,well,moderate,and poor.The contribution of this research is to develop a magnification invariant hybrid feature set comprising cartoon feature,Gabor wavelet,wavelet moments,HSV histogram,color auto-correlogram,color moments,and morphological features that can be used to characterize different grades.Besides,the classifier is modeled as a multiclass structure with six binary class Bayesian optimized random forest(BO-RF)classifiers.This study uses four datasets(two collected from Indian hospitals—Ishita Pathology Center(IPC)of 4X,10X,and 40X and Aster Medcity(AMC)of 10X,20X,and 40X—two benchmark datasets—gland segmentation(GlaS)of 20X and IMEDIATREAT of 10X)comprising multiple microscope magnifications.Experimental results demonstrate that the proposed method outperforms the other methods used for colon cancer grading in terms of accuracy(97.25%-IPC,94.40%-AMC,97.58%-GlaS,99.16%-Imediatreat),sensitivity(0.9725-IPC,0.9440-AMC,0.9807-GlaS,0.9923-Imediatreat),specificity(0.9908-IPC,0.9813-AMC,0.9907-GlaS,0.9971-Imediatreat)and F-score(0.9725-IPC,0.9441-AMC,0.9780-GlaS,0.9923-Imediatreat).The generalizability of the model to any magnified input image is validated by training in one dataset and testing in another dataset,highlighting strong concordance in multiclass classification and evidencing its effective use in the first level of automatic biopsy grading and second opinion.展开更多
The degree of pest damage evaluation on corps in the field environment is very important for precision spraying pesticides.In this paper,we proposed an image processing method to identify the wormholes in the image of...The degree of pest damage evaluation on corps in the field environment is very important for precision spraying pesticides.In this paper,we proposed an image processing method to identify the wormholes in the image of broccoli seedlings,and then to evaluate the damage of the broccoli seedlings by pests.The broccoli seedlings were taken as the research object.The ratio of wormhole areas to broccoli seedling leaves areas(Rw)was used to describe the pest damage degree.An algorithm was developed to calculate the ratio of wormhole areas to broccoli seedling leaves areas.Firstly,broccoli seedling leaves were segmented from the background and the area of the leaves was obtained.There were some holes in segmentation results due to pest damage and other reasons.Then,a classifier based on machine learning was developed to classify the wormholes and other holes.Twenty-four features,including color features and shape features of the holes,were used to develop classifiers.After identifying wormholes from images,the area of the wormholes was obtained and the degree of pest damage to broccoli seedling was calculated.The determination coefficient(R2)between the algorithm calculated pest damage degree and manually labeled pest damage degree was 0.85.The root-mean-square error(d)was 0.02.Results demonstrated that the color and shape were able to effectively segment wormholes from leaves of broccoli seedlings and evaluate the degree of pest damage.This method could provide references for precision spraying pesticides.展开更多
This study explores an automated framework to assist the recognition of hemorrhage traces and bleeding lesions in video streams of small bowel capsule endoscopy (SBCE). The proposed methodology aims to achieve fast im...This study explores an automated framework to assist the recognition of hemorrhage traces and bleeding lesions in video streams of small bowel capsule endoscopy (SBCE). The proposed methodology aims to achieve fast image control (<10 minutes), save valuable time of the physicians, and enable high performance diagnosis. A specialized elimination algorithm excludes all identical consecutive frames by utilizing the difference of gray levels in pixel luminance. An image filtering algorithm is proposed based on an experimentally calculated bleeding index and blood-color chart, which inspects all remaining frames of the footage and identifies pixels that reflect active or potential hemorrhage in color. The bleeding index and blood-color chart are estimated of the chromatic thresholds in RGB and HSV color spaces, and have been extracted after experimenting with more than 3200 training images, derived from 99 videos of a pool of 138 patients. The dataset has been provided by a team of expert gastroenterologist surgeons, who have also evaluated the results. The proposed algorithms are tested on a set of more than 1000 selected frame samples from the entire 39 testing videos, to a prevalence of 50% pathologic frames (balanced dataset). The frame elimination of identical and consecutive frames achieved a reduction of 36% of total frames. The best statistical performance for diagnosis of positive pathological frames from a video stream is achieved by utilizing masks in the HSV color model, with sensitivity up to 99%, precision 94.41% to a prevalence of 50%, accuracy up to 96.1%, FNR 1%, FPR 6.8%. The estimated blood-color chart will be clinically validated and used in feature extraction schemes supporting machine learning ML algorithms to improve the localization potential.展开更多
Recognition and counting of greenhouse pests are important for monitoring and forecasting pest population dynamics.This study used image processing techniques to recognize and count whiteflies and thrips on a sticky t...Recognition and counting of greenhouse pests are important for monitoring and forecasting pest population dynamics.This study used image processing techniques to recognize and count whiteflies and thrips on a sticky trap located in a greenhouse environment.The digital images of sticky traps were collected using an image-acquisition system under different greenhouse conditions.If a single color space is used,it is difficult to segment the small pests correctly because of the detrimental effects of non-uniform illumination in complex scenarios.Therefore,a method that first segments object pests in two color spaces using the Prewitt operator in I component of the hue-saturation-intensity(HSI)color space and the Canny operator in the B component of the Lab color space was proposed.Then,the segmented results for the two-color spaces were summed and achieved 91.57%segmentation accuracy.Next,because different features of pests contribute differently to the classification of pest species,the study extracted multiple features(e.g.,color and shape features)in different color spaces for each segmented pest region to improve the recognition performance.Twenty decision trees were used to form a strong ensemble learning classifier that used a majority voting mechanism and obtains 95.73%recognition accuracy.The proposed method is a feasible and effective way to process greenhouse pest images.The system accurately recognized and counted pests in sticky trap images captured under real greenhouse conditions.展开更多
Citrus Huanglongbing(HLB),which is spread by the citrus psyllid,is the most destructive disease of citrus industry.While no effective cure for the disease has been reported,detection and removal of infected trees can ...Citrus Huanglongbing(HLB),which is spread by the citrus psyllid,is the most destructive disease of citrus industry.While no effective cure for the disease has been reported,detection and removal of infected trees can prevent spreading.Symptoms indicative of HLB can be present in both HLB-positive trees and HLB-negative trees,making identification of infected trees difficult.A detection method for citrus HLB based on image feature extraction and two-stage back propagation neural network(BPNN)modeling was investigated in this research.The identification method for eight different classes including healthy,HLB and non-HLB symptoms was studied.Thirty-four statistical features including color and texture were extracted for each leaf sample,following the two-stage BPNN to model and identify HLB-positive leaves from HLB-negative leaves.The discrimination accuracy can reach approximately 92%which shows that this method based on visual image processing can perform well in detecting citrus HLB.展开更多
In the last few decades,crowd detection has gained much interest from the research community to assist a variety of applications in surveillance systems.While human detection in partially crowded scenarios have achiev...In the last few decades,crowd detection has gained much interest from the research community to assist a variety of applications in surveillance systems.While human detection in partially crowded scenarios have achieved many reliable works,a highly dense crowdlike situation still is far from being solved.Densely crowded scenes offer patterns that could be used to tackle these challenges.This problem is challenging due to the crowd volume,occlusions,clutter and distortion.Crowd region classification is a precursor to several types of applications.In this paper,we propose a novel approach for crowd region detection in outdoor densely crowded scenarios based on color variation context and RGB channel dissimilarity.Experimental results are presented to demonstrate the effectiveness of the new color-based features for better crowd region detection.展开更多
Purpose-The humans are gifted with the potential of recognizing others by their uniqueness,in addition with more other demographic characteristics such as ethnicity(or race),gender and age,respectively.Over the decade...Purpose-The humans are gifted with the potential of recognizing others by their uniqueness,in addition with more other demographic characteristics such as ethnicity(or race),gender and age,respectively.Over the decades,a vast count of researchers had undergone in the field of psychological,biological and cognitive sciences to explore how the human brain characterizes,perceives and memorizes faces.Moreover,certain computational advancements have been developed to accomplish several insights into this issue.Design/methodology/approach-This paper intends to propose a new race detection model using face shape features.The proposed model includes two key phases,namely.(a)feature extraction(b)detection.The feature extraction is the initial stage,where the face color and shape based features get mined.Specifically,maximally stable extremal regions(MSER)and speeded-up robust transform(SURF)are extracted under shape features and dense color feature are extracted as color feature.Since,the extracted features are huge in dimensions;they are alleviated under principle component analysis(PCA)approach,which is the strongest model for solving“curse of dimensionality”.Then,the dimensional reduced features are subjected to deep belief neural network(DBN),where the race gets detected.Further,to make the proposed framework more effective with respect to prediction,the weight of DBNis fine tuned with a new hybrid algorithm referred as lion mutated and updated dragon algorithm(LMUDA),which is the conceptual hybridization of lion algorithm(LA)and dragonfly algorithm(DA).Findings-The performance of proposed work is compared over other state-of-the-art models in terms of accuracy and error performance.Moreover,LMUDA attains high accuracy at 100th iteration with 90%of training,which is 11.1,8.8,5.5 and 3.3%better than the performance when learning percentage(LP)550%,60%,70%,and 80%,respectively.More particularly,the performance of proposed DBNþLMUDAis 22.2,12.5 and 33.3%better than the traditional classifiers DCNN,DBN and LDA,respectively.Originality/value-This paper achieves the objective detecting the human races from the faces.Particularly,MSER feature and SURF features are extracted under shape features and dense color feature are extracted as color feature.As a novelty,to make the race detection more accurate,the weight of DBNis fine tuned with a new hybrid algorithm referred as LMUDA,which is the conceptual hybridization of LA and DA,respectively.展开更多
To explore the correlation between crop leaf digital RGB(Red,Green and Blue)image features and the corresponding moisture content of the leaf,a Canon digital camera was used to collect image information from detached ...To explore the correlation between crop leaf digital RGB(Red,Green and Blue)image features and the corresponding moisture content of the leaf,a Canon digital camera was used to collect image information from detached leaves of heading-stage maize.A drying method was adopted to measure the moisture content of the leaf samples,and image processing technologies,including gray level co-occurrence matrices and grayscale histograms,was used to extract the maize leaf texture feature parameters and color feature parameters.The correlations of these feature parameters with moisture content were analyzed.It is found that the texture parameters of maize leaf RGB images,including contrast,correlation,entropy and energy,were not significantly correlated with moisture content.Thus,it was difficult to use these features to predict moisture content.Of the six groups of eigenvalues for the leaf color feature parameters,including mean,variance,energy,entropy,kurtosis and skewness,mean and kurtosis were found to be correlated with moisture content.Thus,these features could be used to predict the leaf moisture content.The correlation coefficient(R2)of the mean-moisture content relationship model was 0.7017,and the error of the moisture content prediction was within±2%.The R2 of the kurtosis-moisture content relationship model was 0.7175,and the error of the moisture content prediction was within±1.5%.The study results proved that RGB images of crop leaves could be used to measure moisture content.展开更多
In the agriculture field,one of the recent research topics is recognition and classification of diseases from the leaf images of a plant.The recognition of agricultural plant diseases by utilizing the image processing...In the agriculture field,one of the recent research topics is recognition and classification of diseases from the leaf images of a plant.The recognition of agricultural plant diseases by utilizing the image processing techniques will minimize the reliance on the farmers to protect the agricultural products.In this paper,Recognition and Classification of Paddy Leaf Diseases using Optimized Deep Neural Network with Jaya Algorithm is proposed.For the image acquisition the images of rice plant leaves are directly captured from the farm field for normal,bacterial blight,brown spot,sheath rot and blast diseases.In pre-processing,for the background removal the RGB images are converted into HSV images and based on the hue and saturation parts binary images are extracted to split the diseased and non-diseased part.For the segmentation of diseased portion,normal portion and background a clustering method is used.Classification of diseases is carried out by using Optimized Deep Neural Network with Jaya Optimization Algorithm(DNN_JOA).In order to precise the stability of this approach a feedback loop is generated in the post processing step.The experimental results are evaluated and compared with ANN,DAE and DNN.The proposed method achieved high accuracy of 98.9%for the blast affected,95.78%for the bacterial blight,92%for the sheath rot,94%for the brown spot and 90.57%for the normal leaf image.展开更多
Chemical sucker control has been proven to be an effective substitute for manual and mechanical removals.Recognition and location of suckers is the key technology of precision targeted spray which can reduce spray vol...Chemical sucker control has been proven to be an effective substitute for manual and mechanical removals.Recognition and location of suckers is the key technology of precision targeted spray which can reduce spray volume than current spray pattern.The goal of this research was to develop a quick and effective segmentation algorithm of sucker images for real-time mobile targeted spray by evaluating and comparing seven segmentation algorithms categorized into segmentation based on color feature(ExG,ExGExR,and CIVE),K-means clustering segmentation in CIE L*a*b*space(K-Lab),and mean shift clustering segmentation based on color feature(ExG-MS,ExGExR-MS,and CIVE-MS)from time consuming and accuracy.The results indicated that ExGExR and CIVE took shorter time than other algorithms,and were more suitable for real-time operation.By further evaluating segmentation accuracy,ExGExR,CIVE,and mean shift algorithms were acceptable to kill suckers.And ExGExR was the best algorithm for sucker segmentation in consideration of time consuming and accuracy,next came CIVE.展开更多
文摘The transmission of video content over a network raises various issues relating to copyright authenticity,ethics,legality,and privacy.The protection of copyrighted video content is a significant issue in the video industry,and it is essential to find effective solutions to prevent tampering and modification of digital video content during its transmission through digital media.However,there are stillmany unresolved challenges.This paper aims to address those challenges by proposing a new technique for detectingmoving objects in digital videos,which can help prove the credibility of video content by detecting any fake objects inserted by hackers.The proposed technique involves using two methods,the H.264 and the extraction color features methods,to embed and extract watermarks in video frames.The study tested the performance of the system against various attacks and found it to be robust.The evaluation was done using different metrics such as Peak-Signal-to-Noise Ratio(PSNR),Mean Squared Error(MSE),Structural Similarity Index Measure(SSIM),Bit Correction Ratio(BCR),and Normalized Correlation.The accuracy of identifying moving objects was high,ranging from 96.3%to 98.7%.The system was also able to embed a fragile watermark with a success rate of over 93.65%and had an average capacity of hiding of 78.67.The reconstructed video frames had high quality with a PSNR of at least 65.45 dB and SSIMof over 0.97,making them imperceptible to the human eye.The system also had an acceptable average time difference(T=1.227/s)compared with other state-of-the-art methods.
基金The MKE(the Ministry of Knowledge Economy),Korea,under the ITRC(Information Technology Research Center)support program supervised by the NIPA(National IT Industry Promotion Agency) (NIPA-2012-C1090-1121-0010)The Brain Korea21Project in 2012
文摘In order to reduce redundant empty bin capacity in the probability representation,we present a new color feature arrangement mechanism for mean shift tracking objects.In the proposed mechanism,the important optimal color,or we call it optimal color vector,is clustered by closing Euclidean distance which happens inside the original RGB color 3-D spatial domain.After obtaining clustering colors from the reference image RGB spatial domain,novel clustering groups substitute for original color data.So the new color substitution distribution is as similar as the original one.And then target region in the candidate frame is mapped by the constructed optimal clustering colors and the cluster Indices.In the final,mean shift algorithm gives a performance in the new optimal color distribution.Comparison under the same circumstance between the proposed algorithm and conventional mean shift algorithm shows that the former has a certain advantage in computation cost.
基金This work was supported in part by the National Key Research and Development Plan for the 13th Five-Year Plan under Grant 2016YFD0700200This work was supported in part by the National High Technology Research and Development Program of China(2013AA102306).
文摘In the process of tracking the target of the pig,with the change of the size of the tracking target in the video image,the estimated tracking target scale cannot be adaptively updated in real-time,resulting in the low accuracy of the tracking target.In this study,a multi-channel color feature adaptive fusion algorithm was proposed,and the target scale of the pig was updated in real-time by utilizing the contour information of the target pig.Experiments show that the proposed algorithm had a distance precision of 89.7%and an overlap precision of 87.5%,and the average running speed of this algorithm was 50.1 fps.The robustness of the proposed algorithm in tracking target deformation and scale variation were significantly improved,which satisfies the accuracy and real-time requirements of pig target tracking.
基金The work of this paper was supported by the National Natural Science Foundation of China under grant numbers 61572038 received by J.Z.in 2015.URL:https://isisn.nsfc.gov.cn/egrantindex/funcindex/prjsearch-list。
文摘Real-time detection of driver fatigue status is of great significance for road traffic safety.In this paper,a proposed novel driver fatigue detection method is able to detect the driver’s fatigue status around the clock.The driver’s face images were captured by a camera with a colored lens and an infrared lens mounted above the dashboard.The landmarks of the driver’s face were labeled and the eye-area was segmented.By calculating the aspect ratios of the eyes,the duration of eye closure,frequency of blinks and PERCLOS of both colored and infrared,fatigue can be detected.Based on the change of light intensity detected by a photosensitive device,the weight matrix of the colored features and the infrared features was adjusted adaptively to reduce the impact of lighting on fatigue detection.Video samples of the driver’s face were recorded in the test vehicle.After training the classification model,the results showed that our method has high accuracy on driver fatigue detection in both daytime and nighttime.
基金supported by the National Natural Science Foundation of China(61271315)the State Scholarship Fund of China
文摘Image matching based on scale invariant feature transform(SIFT) is one of the most popular image matching algorithms, which exhibits high robustness and accuracy. Grayscale images rather than color images are generally used to get SIFT descriptors in order to reduce the complexity. The regions which have a similar grayscale level but different hues tend to produce wrong matching results in this case. Therefore, the loss of color information may result in decreasing of matching ratio. An image matching algorithm based on SIFT is proposed, which adds a color offset and an exposure offset when converting color images to grayscale images in order to enhance the matching ratio. Experimental results show that the proposed algorithm can effectively differentiate the regions with different colors but the similar grayscale level, and increase the matching ratio of image matching based on SIFT. Furthermore, it does not introduce much complexity than the traditional SIFT.
基金Supported by the Program for Technology Innovation Team of Ningbo Government (No. 2011B81002)the Ningbo University Science Research Foundation (No.xkl11075)
文摘The generic Meanshift is susceptible to interference of background pixels with the target pixels in the kernel of the reference model, which compromises the tracking performance. In this paper, we enhance the target color feature by attenuating the background color within the kernel through enlarging the pixel weightings which map to the pixels on the target. This way, the background pixel interference is largely suppressed in the color histogram in the course of constructing the target reference model. In addition, the proposed method also reduces the number of Meanshift iterations, which speeds up the algorithmic convergence. The two tests validate the proposed approach with improved tracking robustness on real-world video sequences.
基金This work was partially supported by the Research Groups Program(Research Group Number RG-1439-033),under the Deanship of Scientific Research,King Saud University,Riyadh,Saudi Arabia.
文摘Automated grading of colon biopsy images across all magnifications is challenging because of tailored segmentation and dependent features on each magnification.This work presents a novel approach of robust magnification-independent colon cancer grading framework to distinguish colon biopsy images into four classes:normal,well,moderate,and poor.The contribution of this research is to develop a magnification invariant hybrid feature set comprising cartoon feature,Gabor wavelet,wavelet moments,HSV histogram,color auto-correlogram,color moments,and morphological features that can be used to characterize different grades.Besides,the classifier is modeled as a multiclass structure with six binary class Bayesian optimized random forest(BO-RF)classifiers.This study uses four datasets(two collected from Indian hospitals—Ishita Pathology Center(IPC)of 4X,10X,and 40X and Aster Medcity(AMC)of 10X,20X,and 40X—two benchmark datasets—gland segmentation(GlaS)of 20X and IMEDIATREAT of 10X)comprising multiple microscope magnifications.Experimental results demonstrate that the proposed method outperforms the other methods used for colon cancer grading in terms of accuracy(97.25%-IPC,94.40%-AMC,97.58%-GlaS,99.16%-Imediatreat),sensitivity(0.9725-IPC,0.9440-AMC,0.9807-GlaS,0.9923-Imediatreat),specificity(0.9908-IPC,0.9813-AMC,0.9907-GlaS,0.9971-Imediatreat)and F-score(0.9725-IPC,0.9441-AMC,0.9780-GlaS,0.9923-Imediatreat).The generalizability of the model to any magnified input image is validated by training in one dataset and testing in another dataset,highlighting strong concordance in multiclass classification and evidencing its effective use in the first level of automatic biopsy grading and second opinion.
文摘The degree of pest damage evaluation on corps in the field environment is very important for precision spraying pesticides.In this paper,we proposed an image processing method to identify the wormholes in the image of broccoli seedlings,and then to evaluate the damage of the broccoli seedlings by pests.The broccoli seedlings were taken as the research object.The ratio of wormhole areas to broccoli seedling leaves areas(Rw)was used to describe the pest damage degree.An algorithm was developed to calculate the ratio of wormhole areas to broccoli seedling leaves areas.Firstly,broccoli seedling leaves were segmented from the background and the area of the leaves was obtained.There were some holes in segmentation results due to pest damage and other reasons.Then,a classifier based on machine learning was developed to classify the wormholes and other holes.Twenty-four features,including color features and shape features of the holes,were used to develop classifiers.After identifying wormholes from images,the area of the wormholes was obtained and the degree of pest damage to broccoli seedling was calculated.The determination coefficient(R2)between the algorithm calculated pest damage degree and manually labeled pest damage degree was 0.85.The root-mean-square error(d)was 0.02.Results demonstrated that the color and shape were able to effectively segment wormholes from leaves of broccoli seedlings and evaluate the degree of pest damage.This method could provide references for precision spraying pesticides.
文摘This study explores an automated framework to assist the recognition of hemorrhage traces and bleeding lesions in video streams of small bowel capsule endoscopy (SBCE). The proposed methodology aims to achieve fast image control (<10 minutes), save valuable time of the physicians, and enable high performance diagnosis. A specialized elimination algorithm excludes all identical consecutive frames by utilizing the difference of gray levels in pixel luminance. An image filtering algorithm is proposed based on an experimentally calculated bleeding index and blood-color chart, which inspects all remaining frames of the footage and identifies pixels that reflect active or potential hemorrhage in color. The bleeding index and blood-color chart are estimated of the chromatic thresholds in RGB and HSV color spaces, and have been extracted after experimenting with more than 3200 training images, derived from 99 videos of a pool of 138 patients. The dataset has been provided by a team of expert gastroenterologist surgeons, who have also evaluated the results. The proposed algorithms are tested on a set of more than 1000 selected frame samples from the entire 39 testing videos, to a prevalence of 50% pathologic frames (balanced dataset). The frame elimination of identical and consecutive frames achieved a reduction of 36% of total frames. The best statistical performance for diagnosis of positive pathological frames from a video stream is achieved by utilizing masks in the HSV color model, with sensitivity up to 99%, precision 94.41% to a prevalence of 50%, accuracy up to 96.1%, FNR 1%, FPR 6.8%. The estimated blood-color chart will be clinically validated and used in feature extraction schemes supporting machine learning ML algorithms to improve the localization potential.
基金This work was financially supported by the National Natural Science Foundation of China(Grant No.61601034)and the National Natural Science Foundation of China(Grant No.31871525)The authors acknowledge Kimberly Moravec,PhD,from Liwen Bianji,Edanz Editing China(www.liwenbianji.cn/ac),for editing the English text of a draft of this manuscript.
文摘Recognition and counting of greenhouse pests are important for monitoring and forecasting pest population dynamics.This study used image processing techniques to recognize and count whiteflies and thrips on a sticky trap located in a greenhouse environment.The digital images of sticky traps were collected using an image-acquisition system under different greenhouse conditions.If a single color space is used,it is difficult to segment the small pests correctly because of the detrimental effects of non-uniform illumination in complex scenarios.Therefore,a method that first segments object pests in two color spaces using the Prewitt operator in I component of the hue-saturation-intensity(HSI)color space and the Canny operator in the B component of the Lab color space was proposed.Then,the segmented results for the two-color spaces were summed and achieved 91.57%segmentation accuracy.Next,because different features of pests contribute differently to the classification of pest species,the study extracted multiple features(e.g.,color and shape features)in different color spaces for each segmented pest region to improve the recognition performance.Twenty decision trees were used to form a strong ensemble learning classifier that used a majority voting mechanism and obtains 95.73%recognition accuracy.The proposed method is a feasible and effective way to process greenhouse pest images.The system accurately recognized and counted pests in sticky trap images captured under real greenhouse conditions.
基金the supporting of National Natural Science Foundation of China:The research of citrus Huanglongbing in-field detection based on low-altitude multi-sensor fusion(Grant No.61675003)the National Key Research and Development Plan:High efficient ground and aerial spraying technology and intelligent equipment(Grant No.2016YFD0200700).
文摘Citrus Huanglongbing(HLB),which is spread by the citrus psyllid,is the most destructive disease of citrus industry.While no effective cure for the disease has been reported,detection and removal of infected trees can prevent spreading.Symptoms indicative of HLB can be present in both HLB-positive trees and HLB-negative trees,making identification of infected trees difficult.A detection method for citrus HLB based on image feature extraction and two-stage back propagation neural network(BPNN)modeling was investigated in this research.The identification method for eight different classes including healthy,HLB and non-HLB symptoms was studied.Thirty-four statistical features including color and texture were extracted for each leaf sample,following the two-stage BPNN to model and identify HLB-positive leaves from HLB-negative leaves.The discrimination accuracy can reach approximately 92%which shows that this method based on visual image processing can perform well in detecting citrus HLB.
基金the Ministry of Higher Education Malaysia through Fundamental Research Grant Scheme(FRGS)and managed by Universiti Teknologi Malaysia under Vot No.Q.J130000.2508.13491the Machine Learning Research Group+1 种基金Prince Sultan University RiyadhSaudi Arabia[RG-CCIS-2017-06-16].
文摘In the last few decades,crowd detection has gained much interest from the research community to assist a variety of applications in surveillance systems.While human detection in partially crowded scenarios have achieved many reliable works,a highly dense crowdlike situation still is far from being solved.Densely crowded scenes offer patterns that could be used to tackle these challenges.This problem is challenging due to the crowd volume,occlusions,clutter and distortion.Crowd region classification is a precursor to several types of applications.In this paper,we propose a novel approach for crowd region detection in outdoor densely crowded scenarios based on color variation context and RGB channel dissimilarity.Experimental results are presented to demonstrate the effectiveness of the new color-based features for better crowd region detection.
文摘Purpose-The humans are gifted with the potential of recognizing others by their uniqueness,in addition with more other demographic characteristics such as ethnicity(or race),gender and age,respectively.Over the decades,a vast count of researchers had undergone in the field of psychological,biological and cognitive sciences to explore how the human brain characterizes,perceives and memorizes faces.Moreover,certain computational advancements have been developed to accomplish several insights into this issue.Design/methodology/approach-This paper intends to propose a new race detection model using face shape features.The proposed model includes two key phases,namely.(a)feature extraction(b)detection.The feature extraction is the initial stage,where the face color and shape based features get mined.Specifically,maximally stable extremal regions(MSER)and speeded-up robust transform(SURF)are extracted under shape features and dense color feature are extracted as color feature.Since,the extracted features are huge in dimensions;they are alleviated under principle component analysis(PCA)approach,which is the strongest model for solving“curse of dimensionality”.Then,the dimensional reduced features are subjected to deep belief neural network(DBN),where the race gets detected.Further,to make the proposed framework more effective with respect to prediction,the weight of DBNis fine tuned with a new hybrid algorithm referred as lion mutated and updated dragon algorithm(LMUDA),which is the conceptual hybridization of lion algorithm(LA)and dragonfly algorithm(DA).Findings-The performance of proposed work is compared over other state-of-the-art models in terms of accuracy and error performance.Moreover,LMUDA attains high accuracy at 100th iteration with 90%of training,which is 11.1,8.8,5.5 and 3.3%better than the performance when learning percentage(LP)550%,60%,70%,and 80%,respectively.More particularly,the performance of proposed DBNþLMUDAis 22.2,12.5 and 33.3%better than the traditional classifiers DCNN,DBN and LDA,respectively.Originality/value-This paper achieves the objective detecting the human races from the faces.Particularly,MSER feature and SURF features are extracted under shape features and dense color feature are extracted as color feature.As a novelty,to make the race detection more accurate,the weight of DBNis fine tuned with a new hybrid algorithm referred as LMUDA,which is the conceptual hybridization of LA and DA,respectively.
基金This paper was founded by the National Science&Technology Supporting Plan(2012BAH29B04-02)the Science and Technology Innovation Project from Northwest A&F University(Z109021202).
文摘To explore the correlation between crop leaf digital RGB(Red,Green and Blue)image features and the corresponding moisture content of the leaf,a Canon digital camera was used to collect image information from detached leaves of heading-stage maize.A drying method was adopted to measure the moisture content of the leaf samples,and image processing technologies,including gray level co-occurrence matrices and grayscale histograms,was used to extract the maize leaf texture feature parameters and color feature parameters.The correlations of these feature parameters with moisture content were analyzed.It is found that the texture parameters of maize leaf RGB images,including contrast,correlation,entropy and energy,were not significantly correlated with moisture content.Thus,it was difficult to use these features to predict moisture content.Of the six groups of eigenvalues for the leaf color feature parameters,including mean,variance,energy,entropy,kurtosis and skewness,mean and kurtosis were found to be correlated with moisture content.Thus,these features could be used to predict the leaf moisture content.The correlation coefficient(R2)of the mean-moisture content relationship model was 0.7017,and the error of the moisture content prediction was within±2%.The R2 of the kurtosis-moisture content relationship model was 0.7175,and the error of the moisture content prediction was within±1.5%.The study results proved that RGB images of crop leaves could be used to measure moisture content.
文摘In the agriculture field,one of the recent research topics is recognition and classification of diseases from the leaf images of a plant.The recognition of agricultural plant diseases by utilizing the image processing techniques will minimize the reliance on the farmers to protect the agricultural products.In this paper,Recognition and Classification of Paddy Leaf Diseases using Optimized Deep Neural Network with Jaya Algorithm is proposed.For the image acquisition the images of rice plant leaves are directly captured from the farm field for normal,bacterial blight,brown spot,sheath rot and blast diseases.In pre-processing,for the background removal the RGB images are converted into HSV images and based on the hue and saturation parts binary images are extracted to split the diseased and non-diseased part.For the segmentation of diseased portion,normal portion and background a clustering method is used.Classification of diseases is carried out by using Optimized Deep Neural Network with Jaya Optimization Algorithm(DNN_JOA).In order to precise the stability of this approach a feedback loop is generated in the post processing step.The experimental results are evaluated and compared with ANN,DAE and DNN.The proposed method achieved high accuracy of 98.9%for the blast affected,95.78%for the bacterial blight,92%for the sheath rot,94%for the brown spot and 90.57%for the normal leaf image.
基金This research was partially funded by the State Forestry Administration of China project to introduce advanced international forestry science and technology(Grant No.2013-4-02)China Postdoctoral Science Foundation(Grant No.2014M560897)2013 Technology Foundation for Selected Overseas Chinese Scholar,China Ministry of Personnel.
文摘Chemical sucker control has been proven to be an effective substitute for manual and mechanical removals.Recognition and location of suckers is the key technology of precision targeted spray which can reduce spray volume than current spray pattern.The goal of this research was to develop a quick and effective segmentation algorithm of sucker images for real-time mobile targeted spray by evaluating and comparing seven segmentation algorithms categorized into segmentation based on color feature(ExG,ExGExR,and CIVE),K-means clustering segmentation in CIE L*a*b*space(K-Lab),and mean shift clustering segmentation based on color feature(ExG-MS,ExGExR-MS,and CIVE-MS)from time consuming and accuracy.The results indicated that ExGExR and CIVE took shorter time than other algorithms,and were more suitable for real-time operation.By further evaluating segmentation accuracy,ExGExR,CIVE,and mean shift algorithms were acceptable to kill suckers.And ExGExR was the best algorithm for sucker segmentation in consideration of time consuming and accuracy,next came CIVE.