Color Features Of Image Research Articles

Design and testing of infrared diffractive telescope imaging optical systems

Diffractive telescope is an updated imaging technology, it differs from conventional refractive and reflective imaging system, which is based on the principle of diffraction image. However, one of the great challenges of design this optical system is that the diffractive optical element focuses on different wavelengths of light at different points in space, thereby distorting the color characteristics of image. In this paper, we designs a long-wavelength infrared diffractive telescope imaging system with flat surface Fresnel lens and cancels the infrared optical system chromatic aberration by another flat surface Fresnel lens, achieving broadband light (from 8μm to 12μm) to a common focus. At last, the diffuse spot size and MTF function provide diffractive-limited performance.

Efficient decolorization preserving dominant distinctions

Representing color images in grayscale has practical and theoretical importance. Current color-to-gray transformations seldom ensure both quality and efficiency simultaneously in practice. In this paper, we present an efficient global mapping from color to gray while preserving visually dominant features of color images. Our color-to-gray transformation is based on a variant of traditional Difference of Gaussians band-pass filter, which is called luminance filter. The band-pass filter usually has high responses on regions with discriminative colors from their surroundings for certain band. The grayscale is derived from the luminance passing a series of band-pass filters. Our method is linear in the number of pixels, simple to implement and computationally efficient, making it suitable for high resolution images. Experimental results show that our method produces convincing results for a large number of natural and synthetic images.

Measurement of body joint angles for physical therapy based on mean shift tracking using two low cost Kinect images.

Range of motion (ROM) is commonly used to assess a patient's joint function in physical therapy. Because motion capture systems are generally very expensive, physical therapists mostly use simple rulers to measure patients' joint angles in clinical diagnosis, which will suffer from low accuracy, low reliability, and subjective. In this study we used color and depth image feature from two sets of low-cost Microsoft Kinect to reconstruct 3D joint positions, and then calculate moveable joint angles to assess the ROM. A Gaussian background model is first used to segment the human body from the depth images. The 3D coordinates of the joints are reconstructed from both color and depth images. To track the location of joints throughout the sequence more precisely, we adopt the mean shift algorithm to find out the center of voxels upon the joints. The two sets of Kinect are placed three meters away from each other and facing to the subject. The joint moveable angles and the motion data are calculated from the position of joints frame by frame. To verify the results of our system, we take the results from a motion capture system called VICON as golden standard. Our 150 test results showed that the deviation of joint moveable angles between those obtained by VICON and our system is about 4 to 8 degree in six different upper limb exercises, which are acceptable in clinical environment.

ORB feature based web pornographic image recognition

Taken the requirements of web pornographic image recognition both on precision and speed, a pornographic image recognition method based on Oriented FAST and Rotated BRIEF (ORB) is proposed in this paper. The whole recognition process can be divided into two parts: coarse detection and fine detection. Coarse detection can identify the non-pornographic images with no or fewer skin-color regions and facial images quickly. For the remaining images containing much more skin-color regions, fine detection is conducted, which includes three steps: (1) extract ORB descriptors from the skin-color regions and represent the descriptors based on Bag of Words (BOW) model, (2) construct the feature vector combining ORB feature with 72-dimensional Hue, Saturation, Value (HSV) color feature of the whole image, (3) train the classification model using Support Vector Machine (SVM) and apply it for image recognition. The experimental results show that the proposed method can obtain better recognition precision and drastically reduce the average time cost to 1/4 of the method based on Scale Invariant Feature Transform (SIFT).

Rotation and scale invariant hybrid image descriptor and retrieval

Accurate image retrieval is required to index and retrieve large number of images from huge databases. In this paper, an efficient approach is presented to encode the color and textural features of images from the local neighborhood of each pixel. The color features are extracted by quantizing the RGB color space into a single channel with reduced number of shades. The texture information is encoded with structuring patterns generated from the locally structured elements chosen as a basis. Color and textural features are fused together to construct the inherently rotation and scale-invariant hybrid image descriptor (RSHD). This fusion is carried out by extracting textural cues over each shade independently. RSHD has been tested on the Corel dataset and experimental results suggest that RSHD outperforms state-of-the-art descriptors. The performance of the RSHD is promising under rotation and scaling. It can also be effectively used under more complex image transformations.

Semantic Songket Image Search with Cultural Computing of Symbolic Meaning Extraction and Analytical Aggregation of Color and Shape Features

The term "Songket" comes from the Malay word "Sungkit", which means "to hook" or "to gouge". Every motifs names and variations was derived from plants and animals as source of inspiration to create many patterns of songket. Each of songket patterns have a philosophy in form of rhyme that refers to the nature of the sources of songket patterns and that philosophy reflects to the beliefs and values of Malay culture. In this research, we propose a system to facilitate an understanding of songket and the philosophy as a way to conserve Songket culture. We propose a system which is able to collect information in image songket motif variations based on feature extraction methods. On each image songket motif variations, we extracted philosophy of rhyme into impressions, and extracting color features of songket images using a histogram 3D-Color Vector quantization (3D-CVQ), shape feature extraction songket image using HU Moment invariants. Then, we created an image search based on impressions, and impressions search based on image. We use techniques of search based on color, shape and aggregation (combination of colors and shapes). The experiment using impression as query : 1) Result based on color, the average value of true 7.3, total score 41.9, 2) Result based on shape, the average value of true 3, total score 16.4, 3) Result based on aggregation, the average value of true 3, total score 17.4. While based using Image Query : 1) Result based on color, the average precision 95%, 2) Result based on shape, average precision 43.3%, 3) Based aggregation, the average precision 73.3%. From our experiments, it can be concluded that the best search system using query impression and query image is based on the color.Keyword : Image Search, Philosophy, impression, Songket, cultural computing, Feature Extraction, Analytical aggregation.

Automated Color Logo Recognition System based on Shape and Color Features

This paper proposes an automated system for rotation and scale invariant color logo recognition. Colored logo images are recognized using one shape feature namely Moments Invariant and a color feature namely Color Moments. Shape of the logo is modeled using the first two central normalized Hu’s invariant moments while color is modeled using the mean, standard deviation, skewness and kurtosis calculated from the color channels. Each image is scaled and rotated by arbitrary amounts before being submitted to the recognition engine. Classification is done using Manhattan and Euclidean distances. Experimental verification is obtained using a data set of 900 images divided into 75 classes. The proposed approach is highly scalable and robust providing accuracy results comparable to the state of the art. General Terms Image Processing, Pattern Recognition, Color and Shape features.

Color and multiscale texture features from vectorial mathematical morphology

In this paper, we explore an original way to compute texture features for color images in a vector process. Using a dedicated approach for color ordering, we produce a complete framework for color mathematical morphology adapted to human visual system characteristics. Then, morphological multiscale texture features are defined. To understand the texture feature behavior, we present the feature response to basic images variations. Finally, we compare the texture feature performance in front of a classical classification task using Outex database.

Real-time evaluation of polyphenol oxidase (PPO) activity in lychee pericarp based on weighted combination of spectral data and image features as determined by fuzzy neural network

A novel method of using hyperspectral imaging technique with the weighted combination of spectral data and image features by fuzzy neural network (FNN) was proposed for real-time prediction of polyphenol oxidase (PPO) activity in lychee pericarp. Lychee images were obtained by a hyperspectral reflectance imaging system operating in the range of 400–1000nm. A support vector machine-recursive feature elimination (SVM-RFE) algorithm was applied to eliminating variables with no or little information for the prediction from all bands, resulting in a reduced set of optimal wavelengths. Spectral information at the optimal wavelengths and image color features were then used respectively to develop calibration models for the prediction of PPO in pericarp during storage, and the results of two models were compared. In order to improve the prediction accuracy, a decision strategy was developed based on weighted combination of spectral data and image features, in which the weights were determined by FNN for a better estimation of PPO activity. The results showed that the combined decision model was the best among all of the calibration models, with high R2 values of 0.9117 and 0.9072 and low RMSEs of 0.45% and 0.459% for calibration and prediction, respectively. These results demonstrate that the proposed weighted combined decision method has great potential for improving model performance. The proposed technique could be used for a better prediction of other internal and external quality attributes of fruits.

A Novel Approach for Color Image Feature Extraction using Swarm Intelligence

This paper reports research into representing edge detectors, with the help of swarm intelligence [1]. Other gradient-based detectors may not produce the same edge pixels if applied onto the same image. Hence, the need of an edge detector arises, which is precise in detecting edges in majority of the common types of edges. The idea behind Swarm intelligence arises from the insects, bird flocks, fish schools, and wildebeest herds etc. They all have some common features, they move in groups which are having a special behavior. Their coordination is so good and it observed that as some centralized controller dictates all movement. With this background, swarm intelligence relates with ants which may not be very clever individually, but ant colonies are capable of searching, making plans, and optimizing routes to food. Ant colonies are so good at finding the shortest path from one location to another, that an algorithm was developed based on their behavior called as Ant Colony Optimization. Keeping this idea in mind the paper extends previous work of an ant algorithm which is used for edge detection of gray image. Here, ACO is applied for feature extraction and edge pattern detection of a color image. An adaptive threshold histogram method is introduced to the ACO. An image is distributed separately as a combination of three primary colors Red(R), Green (G) and Blue (B). Based on these color intensities three separate histograms are generated and their average mean value is computed which is the threshold value. Thus, a threshold value is generated dynamically every time. General Terms Edge Detection, Feature Extraction, Swarm Intelligence, pheromone updation and evaporation

Multisource Data Ensemble Modeling for Clinker Free Lime Content Estimate in Rotary Kiln Sintering Processes

Clinker free lime (f-CaO) content plays a crucial role in determining the quality of cement. However, the existing methods are mainly based on laboratory analysis and with significant time delays, which makes the closed-loop control of f-CaO content impossible. In this paper, a multisource data ensemble learning-based soft sensor model is developed for online estimation of clinker f-CaO content. To build such a soft sensor model, input flame images, process variables, and the corresponding output f-CaO content data for a rotary cement kiln were collected from No. 2 rotary kiln at Jiuganghongda Cement Plant which produces 2000 tonnes of clinker per day. The raw data were preprocessed to distinguish the flame image regions of interest (ROI) and remove process variable outliers. Three types of flame image ROI features, i.e., color, global configuration, and local configuration features, were then extracted without segmentation. Further, a kernel partial least square technique was applied for extracting the compressed score matrix features from the concatenated flame image features and filtered process variables to avoid high-dimensional, nonlinear, and correlated problems. Feed-forward neural networks with random weights were employed as base learners in our proposed ensemble modeling framework, which aims to enhance the model’s reliability and prediction performance. A total of 157 flame images, the associated process variable data, and the experimentally measured f-CaO content data were used in our experiments. A comparative study on the f-CaO content estimator built by various feature compressed techniques and learner models and robustness analysis were carried out. The results indicate that the proposed multisource data ensemble soft sensor model performs favorably and has good potential in real world applications.

ESTIMATION OF COPPER CONCENTRATE GRADE BASED ON COLOR FEATURES AND LEAST-SQUARES SUPPORT VECTOR REGRESSION

In this paper, a new method based on color features of microscopic image and least-squares support vector regression model (LS-SVR) is proposed for indirect measurement of copper concentrate grade. Red, green and blue (RGB), hue and color vector angle were extracted from color microscopic images of a copper concentrate sample and selected for the comparison. Three different estimation models based on LS-SVR were developed using RGB, hue, and color vector angle, respectively. A comparison of three models was carried out through a validation test. The best model was obtained for the hue giving a running time of 30.243 ms, root mean square error of 0.8644 and correlation coefficient value of 0.9997. The results indicated that the copper concentrate grade could be estimated by the LS- SVR model using the hue as input parameter with a satisfactory accuracy.

Bovines Muzzle Classification Based on Machine Learning Techniques

Abstract Bovines muzzle classification is considered as a biometric classifier to maintain the safety of bovines and guarantee the livestock products. This paper presents two different bovines classifications models using Artificial Neural Network (ANN) and K-Nearest Neighbor Classifier (KNN). The proposed ANN model consists of three phases; pre-processing, feature extraction and classifications. Pre-processing techniques; histogram equalization and mathematical morphology filtering has been used. The ANN model use Segmentation-based Fractal Texture Analysis (SFTA) for extract muzzle features. The proposed KNN model consists of two phases; Expectation Maximization image segmentation and classification. Expectation Maximization image segmentation (EM) depends on extracts bovine image color and texture feature extraction. The experimental result evaluation proves the advancement of KNN model than ANN as it achieves 100% classification accuracy in case of increase number of classification groups to twenty-five compared to 92.76% classification accuracy achieved from ANN classification model.

A class of nonlocal variational problems on a vector bundle for color image local contrast reduction/enhancement

We extend two existing variational models from the Euclidean space to a vector bundle over a Riemannian manifold. The Euclidean models , dedicated to regularize or enhance some color image features, are based on the concept of nonlocal gradient operator acting on a function of the Euclidean space. We then extend these models by generalizing this operator to a vector bundle over a Riemannian manifold with the help of the parallel transport map associated to some class of covariant derivatives. Through the dual formulations of the proposed models, we obtain the expressions of their solutions, which exhibit the functional spaces that describe the image features. Finally, for a well-chosen covariant derivative and its nonlocal extension, the proposed models perform local contrast modification (reduction or enhancement) and experiments show that they preserve more the aspect of the original image than the Euclidean models do while modifying equally its contrast.

A novel RGB Fourier transform-based color space for optical microscopic image processing

This paper presents an IF1F2 color space for describing colors in microscopic images. Unlike the classical color spaces that treat the R, G, and B (RGB) components of a pixel’s color as independent elements, the proposed color space treats RGB as a sampled spectral signal, called a RGB signal. Then, based on the Fourier spectrum analysis of the RGB signal, the 1-D discrete Fourier transform is introduced to describe the color features of microscopic images. K-means clustering experiments on two microscopic image datasets validate the superiority of the proposed IF1F2 color space compared with the classical RGB and HSV (hue, saturation, and value) color spaces.

Matching methods of classic hail echo cores of weather radar

Because the shape, intensity, and texture features of classic hail echo cores are easy to observe, these features are used to examine the matching methods of classic hail echo cores by using template-based and feature-based matching algorithms in this study. We first introduce the shape, intensity, and texture features of classic hail echo cores and the methods of expressing them quantitatively. Template-based and content-based matching algorithms are then used to calculate the similarity distance between classic hail echoes and real-time echoes and to determine the matching patterns between them based on pixels and features. In addition, we verify and analyze the proposed matching methods by using the case data of an extraordinarily severe hailstorm occurring in the Gansu province of China on May 30, 2005. Our research indicates that the identification rates of these two types of matching methods are very high and that the color features of classic hail echo core images are easy to observe. Moreover, V-shaped notches can also be identified and matched by using the texture feature-based matching method under stricter conditions. However, the results rely to a certain degree on the choices of templates, which indicate that the texture features of classic hail echo core images are not very obvious.

Content-based Image Retrieval using Image Partitioning with Color Histogram and Wavelet-based Color Histogram of the Image

paper presents two content-based image retrieval algorithms that are based on image partitioning. The retrieval in the first algorithm is based only on the image color feature represented by the color histogram, while the retrieval in the second one is based on the image color and texture features represented by the color histogram and Haar wavelet transform, respectively. In these algorithms, each image in the database and the query image are divided into 4-equal sized blocks. Color and texture features are extracted for each block. Distances between the blocks of the query image and the blocks of a database image are calculated, then, the similarity between the query image and the database image is calculated by finding the minimum cost matching based on most similar highest priority (MSHP) principle. A CBIR system that implements the proposed algorithms has been developed. To evaluate the effectiveness of the proposed algorithms, experiments have been carried out using different color quantization schemes for three different color spaces (HSV, YIQ and YCbCr) with two similarity measures, namely the Histogram Euclidean Distance and Histogram Intersection Distance. The WANG image database, which contains 1000 general-purpose color images, has been used in the experiments.

Color Extraction and Research of Image Retrieval Based on Wavelet Territory

This paper firstly studies the image color features based on wavelet territory. We introduce a color features’ extract method based on HSI low-frequency subband color features after partition. Firstly, according to the image attention from human eyes, we split the image into sub-blocks. Then extract HSI low-frequency subband color features of each sub-block after wavelet transform, and we can obtain the image color features by weighting. Comparing with traditional histogram method, the experiment results show that the proposed algorithm based on weighted dominant color feature has better retrieval precision.

Depth map super-resolution using stereo-vision-assisted model

Abstract In this paper, we propose a novel Stereo-Vision-Assisted (SVA) model for depth map super-resolution. Given a low-resolution depth map as input, we recover a high-resolution depth map using the registered high-resolution color stereo image pair. First, based on the mutual benefits between raw depth map and features of high resolution color image, we model the relationship with two constraint terms of local and non-local priors and sufficiently explore their complementary nature. Moreover, by considering reliable disparity pixels calculated from stereo matching algorithm, we formulate a stereo disparity regularization term to further reinforce the preservation of fine depth detail. Hence, our SVA objective function includes a non-local prior constraint, a local prior constraint and a stereo disparity prior constraint. In addition, we employ an efficient algorithm to optimize the objective function. Experimental results demonstrate that our approach can obtain superior performance in comparison with other methods.

Simple yet effective color principal and discriminant feature extraction for representing and recognizing color images

In this paper, we investigate the problem of extracting two-dimensional color principal and discriminant features for understanding color images. Specifically, two simple yet effective color image feature extraction criteria, called Color Principal Component Analysis (ColorPCA) and Color Linear Discriminant Analysis (ColorLDA), are proposed for color image analysis. The presented criteria can preserve color and topology information of pixels in images, and extract features directly from color images in an efficient manner by eigen-decomposing a single eigen-problem. In modeling the criteria, color image scatter matrices are defined. Like PCA, LDA and their two-dimensional (2D) extensions, our methods only need to choose the number of projection vectors. More importantly, the matrices to be eigen-decomposed in our criteria have the same size as 2DPCA and 2DLDA that are very efficient. To achieve an orthogonal projection matrix, trace ratio ColorLDA is also presented. We also present the alternative versions of our approaches for feature learning through mining row or column information of the images. Extensive simulations on benchmark datasets are conducted to evaluate our algorithms. The investigated cases demonstrate the effectiveness and efficiency of our techniques, compared with other most related state-of-the-art 1D and 2D criteria.