Kinds Of Image Features Research Articles

Synthesizing realistic high-resolution retina image by style-based generative adversarial network and its utilization

Realistic image synthesis based on deep learning is an invaluable technique for developing high-performance computer aided diagnosis systems while protecting patient privacy. However, training a generative adversarial network (GAN) for image synthesis remains challenging because of the large amounts of data required for training various kinds of image features. This study aims to synthesize retinal images indistinguishable from real images and evaluate the efficacy of the synthesized images having a specific disease for augmenting class imbalanced datasets. The synthesized images were validated via image Turing tests, qualitative analysis by retinal specialists, and quantitative analyses on amounts and signal-to-noise ratios of vessels. The efficacy of synthesized images was verified by deep learning-based classification performance. Turing test shows that accuracy, sensitivity, and specificity of 54.0 ± 12.3%, 71.1 ± 18.8%, and 36.9 ± 25.5%, respectively. Here, sensitivity represents correctness to find real images among real datasets. Vessel amounts and average SNR comparisons show 0.43% and 1.5% difference between real and synthesized images. The classification performance after augmenting synthesized images outperforms every ratio of imbalanced real datasets. Our study shows the realistic retina images were successfully generated with insignificant differences between the real and synthesized images and shows great potential for practical applications.

A Vision-Based Two-Stage Framework for Inferring Physical Properties of the Terrain

The friction and stiffness properties of the terrain are very important pieces of information for mobile robots in motion control, dynamics parameter adjustment, trajectory planning, etc. Inferring the friction and stiffness properties in advance can improve the safety, adaptability and reliability, and reduce the energy consumption of the robot. This paper proposes a vision-based two-stage framework for pre-estimating physical properties of the terrain. We established a field terrain image dataset with weak annotations. A semantic segmentation network that can segment terrains at the pixel level was designed. Given that the same terrain also has different physical properties, we designed two kinds of image features, and we use a decision-making model to realize the mapping from terrain to physical properties. We trained and tested the network comprehensively, and experimented with the complete framework for estimating physical properties. The experimental results show that our framework has good performance.

Feature fusion via Deep Random Forest for facial age estimation

In the last few years, human age estimation from face images attracted the attention of many researchers in computer vision and machine learning fields. This is due to its numerous applications. In this paper, we propose a new architecture for age estimation based on facial images. It is mainly based on a cascade of classification trees ensembles, which are known recently as a Deep Random Forest. Our architecture is composed of two types of DRF. The first type extends and enhances the feature representation of a given facial descriptor. The second type operates on the fused form of all enhanced representations in order to provide a prediction for the age while taking into account the fuzziness property of the human age. While the proposed methodology is able to work with all kinds of image features, the face descriptors adopted in this work used off-the-shelf deep features allowing to retain both the rich deep features and the powerful enhancement and decision provided by the proposed architecture. Experiments conducted on six public databases prove the superiority of the proposed architecture over other state-of-the-art methods.

Yarn-Dyed Fabric Image Retrieval Using Colour Moments and the Perceptual Hash Algorithm

Due to the variety of yarn colours and arrangement, it is a challenging problem to retrieve a yarn-dyed fabric image. In this paper, yarn-dyed fabric samples are captured by the DigiEye system first, and then pattern images of the fabric images captured are simulated by pattern design software based on extracted structure parameters of the yarn-dyed fabric. For the simulated pattern image, an effective algorithm is proposed to retrieve these kinds of images by combining the colour moments method and perceptual hash algorithm. Then the pattern images retrieved are mapped back to the yarn-dyed fabric image so as to realise the yarn-dyed fabric image retrieval. In the algorithm proposed, the colour moments method is adopted to extract the colour features, and the perceptual hash algorithm is utilised to calculate the spatial features of the simulated pattern images. Then the two kinds of image features are used to compute the similarity between the input original image and each target image based on the Euclidean distance and Hamming distance. Relevant images can be retrieved in dependence on the similarity value, which is determined by calculating the optimum weighted value of the colour features’ similarity and spatial features’ similarity. In order to measure the retrieval efficiency of the method proposed, the accuracy rate and retrieval rate of image retrieval were computed in experiments using a PATTERN image database with 300 images. The experimental results show that the average accuracy rate of the method proposed is 85.30% and the retrieval rate - 53.51% when the weighted value of the colour feature similarity is fixed at 0.45 and the spatial feature similarity is 0.55. It is shown that the method presented is effective to retrieve pattern images of yarn-dyed fabric.

Stacked auto-encoder for classification of polarimetric SAR images based on scattering energy

ABSTRACTThis paper proposes a new algorithm, for polarimetric synthetic aperture radar (PolSAR) classification, based on a stacked auto-encoder and scattering energy. Previous approaches to PolSAR classification predominantly consider only the single pixel of distribution of the polarimetric data and scattering characteristics, and ignore other kinds of image features like the relationship of the local pixels. Besides, because of the complexities of PolSAR data, it is difficult to compute the derivatives that are needed for back-propagation in deep-learning classifiers. To overcome these difficulties, we propose a new approach that combines the scattering power and stacks sparse auto-encoder (Scattering SSAE) for PolSAR classification. Firstly, orientation compensation is used to compensate the polarization orientation angle, reducing the impact of polarimetric angle noise. Secondly, Freeman-Durden decomposition is adopted to extract three basic scattering powers: surface, double bounce and volume. Each PolSAR image pixel is transformed into these scattering powers, yielding a new kind of feature from the PolSAR data. Finally, using the three kinds of scattering power as inputs, we combine local spatial information using a patch-based approach, and use a deep learning architecture to achieve classification. We compare our method against several other state-of-the-art methods using ground-truthed test-data, and show that the Scattering SSAE method achieves higher accuracy than other methods on most categories.

Virtual View Generation Based on 3D-Dense-Attentive GAN Networks.

A binocular vision system is a common perception component of an intelligent vehicle. Benefiting from the biomimetic structure, the system is simple and effective. Which are extremely snesitive on external factors, especially missing vision signals. In this paper, a virtual view-generation algorithm based on generative adversarial networks (GAN) is proposed to enhance the robustness of binocular vision systems. The proposed model consists of two parts: generative network and discriminator network. To improve the quality of a virtual view, a generative network structure based on 3D convolutional neural networks (3D-CNN) and attentive mechanisms is introduced to extract the time-series features from image sequences. To avoid gradient vanish during training, the dense block structure is utilized to improve the discriminator network. Meanwhile, three kinds of image features, including image edge, depth map and optical flow are extracted to constrain the supervised training of model. The final results on KITTI and Cityscapes datasets demonstrate that our algorithm outperforms conventional methods, and the missing vision signal can be replaced by a generated virtual view.

Medical Image Blind Integrity Verification with Krawtchouk Moments

A new blind integrity verification method for medical image is proposed in this paper. It is based on a new kind of image features, known as Krawtchouk moments, which we use to distinguish the original images from the modified ones. Basically, with our scheme, image integrity verification is accomplished by classifying images into the original and modified categories. Experiments conducted on medical images issued from different modalities verified the validity of the proposed method and demonstrated that it can be used to detect and discriminate image modifications of different types with high accuracy. We also compared the performance of our scheme with a state-of-the-art solution suggested for medical images—solution that is based on histogram statistical properties of reorganized block-based Tchebichef moments. Conducted tests proved the better behavior of our image feature set.

Pairwise comparison and rank learning for image quality assessment

To know what kinds of image features are crucial for image quality assessment (IQA) and how these features affect the human visual system (HVS) is still largely beyond human knowledge. Hence, machine learning (ML) is employed to build IQA by simulating the HVS behavior in IQA processes. Support vector machine/regression (SVM/SVR) is a major member of ML. It has been successfully applied to IQA recently. As to image quality rating, the human’s opinion about it is not always reliable. In fact, the subjects cannot precisely rate the small difference of image quality in subjective testing, resulting in unreliable Mean Opinion Scores (MOSs). However, they can easily identify the better/worse one from two given images, even their qualities do not differ much. In this sense, the human’s opinion on pairwise comparison (PC) of image quality is more reliable than image quality rating. Thus, PC has been exploited in developing IQA metrics. In this paper, a rank learning optimization framework is firstly developed to model IQA. Particularly, the PCs of image quality instead of numerical ratings are incorporated into the optimization framework. Then, a novel no-reference (NR)-IQA is proposed to infer image quality in terms of image quality ranks. By importing rank learning theory and PC into IQA, a fundamental and meaningful departure from the existing framework of IQA could be expected. The experimental results confirm that the proposed Pairwise Rank Learning based Image Quality Metric (PRLIQM) can achieve comparable performance over the state-of-the-art NR-IQA approaches.

Ordinal Distance Metric Learning for Image Ranking

Recently, distance metric learning (DML) has attracted much attention in image retrieval, but most previous methods only work for image classification and clustering tasks. In this brief, we focus on designing ordinal DML algorithms for image ranking tasks, by which the rank levels among the images can be well measured. We first present a linear ordinal Mahalanobis DML model that tries to preserve both the local geometry information and the ordinal relationship of the data. Then, we develop a nonlinear DML method by kernelizing the above model, considering of real-world image data with nonlinear structures. To further improve the ranking performance, we finally derive a multiple kernel DML approach inspired by the idea of multiple-kernel learning that performs different kernel operators on different kinds of image features. Extensive experiments on four benchmarks demonstrate the power of the proposed algorithms against some related state-of-the-art methods.

Research on Industry Image Technology with the Integration Algorithm of Color Features and Feedback of Image Retrieval

Current image retrieval methods use only one kind of image features, which can not describe image content completely. In this paper, an image retrieval method based on color and texture integration is proposed. Calculation of similarity for the individual features is normalized before color feature integration. In order to improve the precision of image retrieval, a relevance feedback mechanism is invoked. The experiment results show that the proposed method has an excellent retrieval performance.

An effective method for identifying small objects on a complicated background

This paper deals with automatic object recognition (AOR) on a complex background. Based on the domain knowledge and principles of perceptual organization, a heuristic two-stage identifying algorithm is proposed to extract and identify small objects of unknown location in outdoor images of visible spectral bands. Corresponding heuristic knowledge and assumptions are used for each stage. This algorithm is capable of making effective use of two kinds of image features that are imperfect but have a certain degree of complementarity, i.e. regions and edges, thus enhancing the ability of the machine vision system to extract and identify 3D small objects on a complicated background. Experimental results have verified the effectiveness and practicality of the method proposed.

Neural network vowel-recognition jointly using voice features and mouth shape image

This paper describes a neural approach intended to improve the performance of an automatic speech recognition system for unrestricted speakers by using not only voice sound features but also image features of the mouth shape. In particular, we used the natural sample voice signals and mouth shape images that were acquired in the general environment, neither in the sound isolation room nor under specific lighting conditions. The FFT power spectrum of acoustic speech was used as the voice feature. In addition, the gray level image, binary image and geometrical shape features of the mouth were used as the compensatory information, and compared which kinds of image features are effective. For unrestricted speakers, a vowel recognition rate of about 80% was obtained using only voice features, but this increased to some 92% when voice features plus binary images were used. This method can be applied not only to the improvement of voice recognition, but also to aid the communication of hearing-impaired people.

Speaker‐independent vowel recognition combining voice features and mouth shape image with neural network

AbstractThis paper describes a neural approach intended to improve the performance of a voice recognition system for unrestricted speakers using not only voice sound features but also image features of the mouth shape. The FFT power spectrum of acoustic speech was used as the voice feature. In addition, the gray‐level image, binary image, and geometrical shape features of the mouth were used as the compensatory information and a comparison made of which kinds of image features are effective for voice recognition by a neural network.For unrestricted speakers, a vowel recognition rate of about 80 percent was obtained using only voice features. However, this increased to some 92 percent when voice features plus binary images were used. This method can be applied not only to the improvement of voice recognition, but also to aid the communication of hearing‐impaired people.