Rotation Invariant Features Research Articles

Noise robust and rotation invariant entropy features for texture classification

In this paper, a new formula is proposed that uses local entropy for texture feature extraction. This new method is similar to entropy; however, it calculates the local entropy of each local patch of textures. Entropy (ENT) is an attribute that measures the randomness of gray-level distribution of image. Entropy extracts dissimilarity of each local patch. In this paper, local entropy is compared to Local Binary Pattern (LBP) and local variance (VAR). All of these descriptors are rotation invariant and are used for extracting the features from each local neighborhood of textures. In spite of low accuracy of VAR and LBP the performance of ENT does not decrease significantly for noisy textures. In other words, ENT is more robust to noise than VAR and LBP. Implementations on Outex, UIUC, CUReT and MeasTex datasets show that entropy is more accurate than variance and LBP. Similar to VAR and LBP, ENT can be combined with other descriptors to improve the performance of classification. For almost all datasets that are used in implementation part, LBP/ENT is more accurate than LBP/VAR for normal and noisy textures. Also the ENT accuracy outperforms the accuracy of VAR and LBP and most of the advanced noise robust LBP versions for low Signal to Noise Ratio (SNR) values (SNRź<ź10). ENT feature is a continuous value so it is necessary to quantize to discrete value for histogram. The quantization and train step of ENT is the same as VAR.

Invariant Hough random ferns for RGB-D-based object detection

This paper studies the challenging problem of object detection using rich image and depth features. An invariant Hough random ferns framework for RGB-D images is proposed here, which primarily consists of a rotation-invariant RGB-D local binary feature, random ferns classifier training, Hough mapping and voting, searches for the maxima, and back projection. In comparison with traditional three-dimensional local feature extraction techniques, this method is effective in reducing the amount of computation required for feature extraction and matching. Moreover, the detection results showed that the proposed method is robust against rotation and scale variations, changes in illumination, and part-occlusions. The authors believe that this method will facilitate the use of perception in fields such as robotics.

A new Global-Gabor-Zernike feature descriptor and its application to face recognition

Face recognition is an important subject in computer vision and authentication systems. Feature extraction is one of the main steps in the face recognition systems, which greatly affects recognition accuracy. In the most of the existing methods, only local features in the facial area are extracted and employed in recognizing the person’s face. In this article, at first a novel multi-scale and rotation invariant global feature descriptor is introduced by applying the Zernike moment on the outputs of Gabor filters. Then the proposed global feature along with an efficient local feature, the histogram of oriented gradient (HOG), is employed to propose a new face recognition system. The proposed system was tested on three famous face recognition databases, namely ORL, Yale and AR and face recognition rates of 98%, 97.8% and 97.1% were obtained respectively. These rates are higher than other state-of-the-art methods.

Rotation invariant HOG for object localization in web images

To localize objects in Web images using an invariant descriptor is crucial. The HOG (histogram of oriented gradients) descriptor is used to increase the accuracy of localization. It is a shape descriptor that considers frequencies of gradient orientation in localized portions of an image. This well known descriptor does not cover rotation variations of an object in images. This paper introduces a rotation invariant feature descriptor based on HOG. The proposed descriptor is used in a top-down searching technique that covers the scale variation of the objects in images. The efficiency of this method is validated by comparing the performance with existing research in a similar domain on the Caltech-256 Web dataset. The proposed method not only provides robustness against geometrical transformations of objects but also is computationally more efficient.

Rotation invariant texture classification using extended wavelet channel combining and LL channel filter bank

In this paper, a simple and effective method for rotation invariant texture classification is proposed. The suggested feature vector is small as compared to other state-of-the-art methods. The proposed method is the development of wavelet combining channels approaches that have been introduced for extraction of rotation invariant features. These methods have two big problems with textures with strong directionality and textures with similar spatial frequency and different directionality. Therefore, these methods have not attracted much attention for texture classification. Our method introduces several extensions on the wavelet combining channels idea: first, it changes the combining scheme, second, it applies a rescaling scheme, third, it changes the method of calculating the energy features and forth, it uses low horizontal and low vertical frequency (LL) channel features in each level of decomposition. We also apply a filter bank on the first level of decomposition for extracting the edge features and different texture features. To illustrate the impact of these developments, we applied the proposed method on four different databases, from the Outex, Brodatz, and VisTex, consisting of 24, 25, 54, and 87 different classes of textures, respectively. The results show that the proposed method has comparable and even better performance than the other state-of-the-art methods. The results show that by increasing the number of classes from 25 to 87, maximum classification accuracy has dropped by only 2.4%. For more comparisons, the method is applied on two different illumination textures and it obtains an average of almost 10.9% increase in accuracy than the traditional wavelet channel combining method.

XQ-NLM: Denoising Diffusion MRI Data via x-q Space Non-Local Patch Matching.

Noise is a major issue influencing quantitative analysis in diffusion MRI. The effects of noise can be reduced by repeated acquisitions, but this leads to long acquisition times that can be unrealistic in clinical settings. For this reason, post-acquisition denoising methods have been widely used to improve SNR. Among existing methods, non-local means (NLM) has been shown to produce good image quality with edge preservation. However, currently the application of NLM to diffusion MRI has been mostly focused on the spatial space (i.e., the x-space), despite the fact that diffusion data live in a combined space consisting of the x-space and the q-space (i.e., the space of wavevectors). In this paper, we propose to extend NLM to both x-space and q-space. We show how patch-matching, as required in NLM, can be performed concurrently in x-q space with the help of azimuthal equidistant projection and rotation invariant features. Extensive experiments on both synthetic and real data confirm that the proposed x-q space NLM (XQ-NLM) outperforms the classic NLM.

Human age classification using appearance features and artificial neural network

This paper presents a novel method for human age classification via face images by a computer. The proposed method classifies the human face images into four age groups: child, young, adult and senior adult by using appearance features as ageing features and artificial neural network (ANN) as age classifier. The appearance features consist of both shape and textural features. Only two geometric ratios in combination with newly introduced rotation, scale and translation invariant efficient feature face angle are used as shape features. Local binary pattern histogram (LBPH) of regions of interest in face images are used as textural features. The ANN is designed by using two layer feedforward backpropagation neural networks. The performance of proposed age classification system is evaluated on face images from FG-NET ageing database and achieved greatly improved accuracy of 91.09% and 88.18% for male and female, respectively.

Human age classification using appearance features and artificial neural network

This paper presents a novel method for human age classification via face images by a computer. The proposed method classifies the human face images into four age groups: child, young, adult and senior adult by using appearance features as ageing features and artificial neural network (ANN) as age classifier. The appearance features consist of both shape and textural features. Only two geometric ratios in combination with newly introduced rotation, scale and translation invariant efficient feature face angle are used as shape features. Local binary pattern histogram (LBPH) of regions of interest in face images are used as textural features. The ANN is designed by using two layer feedforward backpropagation neural networks. The performance of proposed age classification system is evaluated on face images from FG-NET ageing database and achieved greatly improved accuracy of 91.09% and 88.18% for male and female, respectively.

CBDF: Compressed Binary Discriminative Feature

Keypoint descriptors play an indisputable role in modern vision, whether it is for image registration, content-based retrieval, 3D reconstruction, or object recognition. In that perspective, SIFT descriptor is the most widely used due to its robustness to scale, rotation, and perspective distortions as well as illumination changes. However, the processing effort and the amount of memory required by SIFT make it ill-suited for applications with limited memory and processing power.In this paper, we propose a novel keypoint descriptor called CBDF (Compressed Binary Discriminative Feature) designed to be accurate, memory efficient and fast at the same time. High speed and memory efficiency are made possible due to a binary descriptor obtained by aggregating gradient features. Local discriminative analysis (LDA) is used to make sure the aggregated feature leads to a discriminative descriptor. Accuracy is also enforced by the use of scale, rotation, and illumination invariant features. The CBDF descriptor is stored on a 32-byte vector (which is one order of magnitude smaller than the usual 512-byte vector for SIFT) while matching is done with the Hamming distance.Extensive evaluations on various benchmark datasets show that CBDF is not only faster and generally more accurate than SIFT and SURF, but also more accurate than recent descriptors such as ORB, BRISK and BRIEF.

Robust Image Hashing Based on Selective Quaternion Invariance

Robust image hashing, which maps the perceptual contents of image to a short digest, is a key technique for tackling the challenges of content-based indexing, searching and copyright protection. In this letter, we propose a quaternion invariance based hashing algorithm that can fuse complementary visual features to compact hash. The proposed algorithm leverages quaternion polar cosine transform to holistically capture the spatial and chromatic characteristics of digital image, and rotation-invariant features are derived from the phase information in the quaternion frequency domain. In addition, we also propose an information theoretic based metric for feature quality assessment and a greedy strategy to select the optimal subset of features for hash computation. Extensive experiments over a large database demonstrate that the proposed work shows higher content identification accuracy than most competing algorithms, and the resulting hash is compact and easy to compute.

Orthogonal Polar V Transforms and application to shape retrieval

The traditional orthogonal moments (e.g., Zernike moments) are formulated with polynomials as their basis that often face the problem of computation difficulty especially with the high-order moments. In this paper, we present a novel set of transforms namely the Polar V Transforms (PVTs). We can use the PVTs not only to generate the rotation-invariant features but also to capture global and local information of images. Since the PVTs basis functions can keep a low order of polynomials, we can significantly speed-up the runtime for computing the kernels. The experimental results have demonstrated that our proposed method outperforms the previous methods in runtimes and achieves very good results in shape retrieval compared to the previous methods especially when the images with high degree of perspective distortions.

RIMOC, a feature to discriminate unstructured motions: Application to violence detection for video-surveillance

In video-surveillance, violent event detection is of utmost interest. Although action recognition has been well studied in computer vision, literature for violence detection in video is far sparser, and even more for surveillance applications. As aggressive events are difficult to define due to their variability and often need high-level interpretation, we decided to first try to characterize what is frequently present in video with violent human behaviors, at a low level: jerky and unstructured motion. Thus, a novel problem-specific Rotation-Invariant feature modeling MOtion Coherence (RIMOC) was proposed, in order to capture its structure and discriminate the unstructured motions. It is based on the eigenvalues obtained from the second-order statistics of the Histograms of Optical Flow vectors from consecutive temporal instants, locally and densely computed, and further embedded into a spheric Riemannian manifold.The proposed RIMOC feature is used to learn statistical models of normal coherent motions in a weakly supervised manner. A multi-scale scheme applied on an inference-based method allows the events with erratic motion to be detected in space and time, as good candidates of aggressive events.We experimentally show that the proposed method produces results comparable to a state-of-the-art supervised approach, with added simplicity in training and computation. Thanks to the compactness of the feature, real-time computation is achieved in learning as well as in detection phase. Extensive experimental tests on more than 18 h of video are provided in different in-lab and real contexts, such as railway cars equipped with on-board cameras.

Efficient Dense-Field Copy–Move Forgery Detection

We propose a new algorithm for the accurate detection and localization of copy–move forgeries, based on rotation-invariant features computed densely on the image. Dense-field techniques proposed in the literature guarantee a superior performance with respect to their keypoint-based counterparts, at the price of a much higher processing time, mostly due to the feature matching phase. To overcome this limitation, we resort here to a fast approximate nearest-neighbor search algorithm, PatchMatch, especially suited for the computation of dense fields over images. We adapt the matching algorithm to deal efficiently with invariant features, so as to achieve higher robustness with respect to rotations and scale changes. Moreover, leveraging on the smoothness of the output field, we implement a simplified and reliable postprocessing procedure. The experimental analysis, conducted on databases available online, proves the proposed technique to be at least as accurate, generally more robust, and typically much faster than the state-of-the-art dense-field references.

Rotation invariant multi-frame image super resolution reconstruction using Pseudo Zernike Moments

The purpose of multi-frame super resolution (SR) is to combine multiple low resolution (LR) images to produce one high resolution (HR) image. The major challenge of classic SR approaches is accurate motion estimation between the frames. To address this problem, fuzzy motion estimation method has been proposed that replaces value of each pixel using the weighted average of all its neighboring pixels in all LR images. However, in case of rotation between LR images, comparing the gray level of blocks is not a suitable criterion for calculating the weight. Hence, magnitude of Zernike Moments (ZM) has been used as a rotation invariant feature. Considering the more robustness of Pseudo Zernike Moments (PZM) to noise and its higher description capability for the same order compared to ZM, in this paper, we propose a new method based on the magnitude of PZM as a rotation invariant descriptor for representing the pixels in the weight calculation. Also, due to the fact that the phase of PZM provides significant information for image reconstruction, we propose a new phase-based PZM descriptor for SR by making the phase coefficients invariant to rotation. Experimental results on several image sequences demonstrate that the proposed algorithm outperforms other currently popular SR techniques from the viewpoint of PSNR, SSIM and visual image quality.

Enhanced Feature Based Mosaicing Technique for Visually and Geometrically Degraded Airborne Synthetic Aperture Radar Images

In airborne synthetic aperture radar (SAR), there was a major problem encountered in the area of image mosaic in the absence of platform information and sensor information (geocoding), when SAR is applied in large-scale scene and the platform faces large changes. In order to enhance real-time performance and robustness of image mosaic, enhancement based Speeded-Up Robust Features (SURF) mosaic method for airborne SAR is proposed in this paper. SURF is a novel scale-invariant and rotation-invariant feature. It is perfect in its high computation, speed and robustness. In this paper, When the SAR image is acquired, initially the image is enhanced by using local statistic techniques and SURF is applied for SAR image matching accord to its characteristic, and then acquires its invariant feature for matching. In the process of image matching, the nearest neighbor rule for initial matching is used, and the wrong points of the matches are removed through RANSAC fitting algorithm. The proposed algorithm is implemented in different SAR images with difference in scale change, rotation change and noise. The proposed algorithm is compared with other existing algorithms and the quantitative and qualitative measures are calculated and tabulated. The proposed algorithm is robust to changes and the threshold is varied accordingly to increase the matching rate more than 95 %.

Handwritten Courtesy Amount and Signature Recognition on Bank Cheque using Neural Network

Bank cheques are used not only in our country but also all over the world for financial transactions. Still now Bank cheques are processed manually everyday in both developed and developing countries. The proposed system presents a recognition system of both handwritten courtesy amount and signature .To read the courtesy amount and signature, the system uses the scanned image of Bank cheque. The proposed system divided into several stages that focus on: Image preprocessing; the detection of block of courtesy amount and signature; the post processing; the segmentation of string into characters; Feature extraction of courtesy amount and signature; Neural Network recognition. At first, scanned image is converted into gray image and then it is filtered. Then detection of the courtesy amount and signature is performed using image cropping method. The cropped images are post-processed to ensure correct recognition. Then courtesy amount is segmented using the segmentation method. The segmentation module has been implemented as a recursive process. Then segmented digit and binary image of signature are passed through the feature extraction process. A rotation, translation, scaling and orientation invariant feature extraction method has been used to extract the features of the input images based on moment feature extraction method. And finally, a back propagation learning algorithm is used to train up the network and tested the performance. Overall success rate of the system is tested in different sorts of numeral and the experimental result shows satisfactory performance.

Rotation-Invariant Object Detection in Remote Sensing Images Based on Radial-Gradient Angle

To improve the detection precision in complicated backgrounds, a novel rotation-invariant object detection method to detect objects in remote sensing images is proposed in this letter. First, a rotation-invariant feature called radial-gradient angle (RGA) is defined and used to find potential object pixels from the detected image blocks by combining with radial distance. Then, a principal direction voting process is proposed to gather the evidence of objects from potential object pixels. Since the RGA combined with the radial distance is discriminative and the voting process gathers the evidence of objects independently, the interference of the backgrounds is effectively reduced. Experimental results demonstrate that the proposed method outperforms other existing well-known methods (such as the shape context-based method and rotation-invariant part-based model) and achieves higher detection precision for objects with different directions and shapes in complicated background. Moreover, the antinoise performance and parameter influence are also discussed.

Noise and rotation invariant RDF descriptor for palmprint identification

Rotation and noise invariant feature extraction is a challenge in palmprint recognition. This work presents a novel RDF descriptor based on Radon, Dual tree complex wavelet, and Fourier transforms. Combined properties of these transforms help to explore efficiency and robustness of RDF descriptor for palmprint identification. Radon transform can capture directional features of the palmprint and is robust to additive white Gaussian noise also. It converts rotation into translation. 1D Dual tree complex wavelet transform (DTCWT) applied on Radon coefficients in angle direction removes translation in Radon coefficients due to palmprint rotation. The magnitude of 2D Fourier transform performed on resultant coefficients helps to extract rotation and illumination invariant features. The performance of the proposed RDF descriptor is evaluated on noisy and rotated palmprints upto 10?. Trained with normal palmprints only, the proposed system gives good results for rotated and noisy palmprints. Experiments are performed on PolyU 2D, CASIA, and IIITDMJ databases. Theoretical foundations and experimental results show the robustness of RDF descriptor against additive white noise and rotation.

Object classification in 3D baggage security computed tomography imagery using visual codebooks

We investigate the performance of a Bag of (Visual) Words (BoW) object classification model as an approach for automated threat object detection within 3D Computed Tomography (CT) imagery from a baggage security context. This poses a novel and unique challenge for rigid object classification within complex and cluttered volumetric imagery. Within this context it extends the BoW model to 3D transmission imagery (X-ray CT) from its conventional application in 2D reflectance (photographic) imagery. We explore combinations of four 3D feature descriptors (Density Histogram (DH), Density Gradient Histogram (DGH), Scale Invariant Feature Transform (SIFT) and Rotation Invariant Feature Transform (RIFT)), three codebook assignment methodologies (hard, kernel and uncertainty) and seven codebook sizes. Optimal performance is achieved using the DH and DGH descriptors in conjunction with an uncertainty assignment methodology. Successful detection rates in excess of 97% for handguns and 89% for bottles and false-positive rates of approximately 2–3% are achieved. We demonstrate that the underlying imaging modality and the irrelevance of illumination and scale invariance within the transmission imagery context considered here result in the favourable performance of simpler density histogram descriptors (DH, DGH) over 3D extensions of the well-established SIFT and RIFT feature descriptor approaches.

Face Recognition by PCA and Improved LBP Fusion Algorithm

This paper proposed a theoretically efficient approach for face recognition based on principal component analysis (PCA) and rotation invariant uniform local binary pattern texture features in order to weaken the effects of varying illumination conditions and facial expressions. Firstly, the rotation invariant uniform LBP operator was adopted to extract the local texture feature of the face images. Then PCA method was used to reduce the dimensionality of the extracted feature and get the eigenfaces. Finally, the nearest distance classification was used to distinguish each face. The method has been accessed on Yale and ATR-Jaffe face databases. Results demonstrate that the proposed method is superior to standard PCA and its recognition rate is higher than the traditional PCA. And the proposed algorithm has strong robustness against the illumination changes, pose, rotation and expressions.