Multi Images Research Articles

MF-LRTC: Multi-filters guided low-rank tensor coding for image restoration

Abstract Image prior information is a determinative factor to tackling with the ill-posed problem. In this paper, we present multi-filters guided low-rank tensor coding (MF-LRTC) model for image restoration. The appeal of constructing a low-rank tensor is obvious in many cases for data that naturally comes from different scales and directions. The MF-LRTC takes advantages of the low-rank tensor coding to capture the sparse convolutional features generated by multi-filters representation. Using such a low-rank tensor coding would reduce the redundancy between feature vectors at neighboring locations and improve the efficiency of the overall sparse representation. In this work, we are committed to achieving this goal by convoluting the target image with Filed-of-Experts (FoE) filters to formulate multi-feature images. Then similarity-grouped cube set extracted from the multi-features images is regarded as a low-rank tensor. The resulting non-convex model is addressed by an efficient ADMM technique. The potential effectiveness of this tensor construction strategy is demonstrated in image restoration including image deblurring and compressed sensing (CS) applications.

Stack-Captioning: Coarse-to-Fine Learning for Image Captioning

The existing image captioning approaches typically train a one-stage sentence decoder, which is difficult to generate rich fine-grained descriptions. On the other hand, multi-stage image caption model is hard to train due to the vanishing gradient problem. In this paper, we propose a coarse-to-fine multi-stage prediction framework for image captioning, composed of multiple decoders each of which operates on the output of the previous stage, producing increasingly refined image descriptions. Our proposed learning approach addresses the difficulty of vanishing gradients during training by providing a learning objective function that enforces intermediate supervisions. Particularly, we optimize our model with a reinforcement learning approach which utilizes the output of each intermediate decoder's test-time inference algorithm as well as the output of its preceding decoder to normalize the rewards, which simultaneously solves the well-known exposure bias problem and the loss-evaluation mismatch problem. We extensively evaluate the proposed approach on MSCOCO and show that our approach can achieve the state-of-the-art performance.

Multi secret sharing scheme using iterative method

Multi secret sharing is used to hide more than one secret information. In this approach secret information is divided in many parts and distributed among participants. Secret information may be in form of images also. Initially, Visual Cryptography (VC) applied for single secret images. After encryption process of VC, encrypted meaningless images are generated known as shares. Nowadays, multiple secret images may be secured by VC also. However, multi secret sharing schemes provide security to multiple images. Obtaining multiple images by using minimum number of shares is really an existing problem. Existing algorithms are limited to fix number of secret images with minimum number of shares. In this paper, we proposed an (2, n, m) multi secret sharing scheme in which exactly n share are generated to encode m secret images. If set of shares taken from qualified set then secret images revealed on the other hand set of shares taken from forbidden set than no information revealed. Minimum two shares are required to obtain one secret images and n or m+1 shares are used to obtain m multi secret images. Perfect alignment of shares between shares is not required to obtain secret image in the proposed technique.

Storage and Data Handling Multi and Hyperspectral Satellite Images based on Apache Parquet

Storage and Data Handling Multi and Hyperspectral Satellite Images based on Apache Parquet

基于多图像融合的MEMS显微三维形貌重构

基于多图像融合的MEMS显微三维形貌重构

Analysis of Pixel Merging for Multi Image Integration for Security Enhancement

Analysis of Pixel Merging for Multi Image Integration for Security Enhancement

Review of Image Fusion and its techniques

This paper shows that image fusion is a energetic area in digital image processing. The main goal of image fusion is always to merge information from multiple images of the exact same view in order to deliver only the useful information. It discuss the different image fusion techniques by utilizing these techniques original images were decomposed into low frequency sub band coefficients and band pass direction sub band coefficients based on the non-sub-sampled contour let transform whereas various maps of visual salient features are created by visual salient features the local energy, the contrast as well as the gradient .Low-frequency sub band coefficients are got by utilizing these visual saliency maps. The literature survey is conducted on various recent techniques of image fusion. The survey shows that although present NSCT turns based fusion outperforms over accessible techniques.

Optimum laplacian wavelet mask based medical image using hybrid cuckoo search – grey wolf optimization algorithm

Multi scale masking techniques are well known in the field of multi modal medical image fusion. In medical image fusion the quality of complement information is important .In multi modal medical image fusion discrete wavelet transform (db4) based techniques provides greater level of approximation but the edge features available is less. The laplacian filter based techniques provides grater edge features. In this paper we propose an Optimum Laplacian Wavelet Mask (OLWM) based fusion using Hybrid Cuckoo Search -Grey Wolf Optimization (HCS-GWO) for multi modal medical image fusion. The HCS-GWO can automatically select the control parameters of grey wolf algorithm using cuckoo search parameters. First, the proposed fusion approach is validated for MR-SPECT, MR-PET, MR-CT and MR T1-T2 image fusion using various fusion evaluation indexes. Later, the conventional grey wolf optimization is modified with cuckoo search algorithm. Experimental results are analyzed using various performance metrics and our OLWM shows improved results than other conventional decomposition techniques.

Mapping Thermokarst Lakes on the Qinghai–Tibet Plateau Using Nonlocal Active Contours in Chinese GaoFen-2 Multispectral Imagery

In order to monitor the response of thermokarst lakes on the Qinghai–Tibet Plateau (QTP) to rapid climatic changes and human activities, an automated method for extracting shorelines from Chinese GaoFen-2 (GF-2) imagery is proposed. First, the water index (WI) images and the potential lake areas are calculated from the preprocessed multispectral imagery and digital elevation model data, respectively. Second, the initial segmentation obtained by global thresholding of the WI images and masking in the potential lake areas are used to implement the contour initialization of active contours models efficiently. Finally, the nonlocal active contours (NLAC) approach is applied to refine the initial segmentation of the WI images, and the final shoreline vector files are produced by some simple and automatic postprocessing steps. Experiments on the GF-2 imagery demonstrate that 1) by exploiting the capability of WI to locate the approximate shoreline effectively around the evolving contour, the processing time of the proposed method can be saved significantly; 2) the NLAC approach can efficiently identify the shoreline by integrating the nonlocal interactions between pairs of patches inside and outside the lake; and 3) the proposed method can conveniently adapt to the multitemporal and multifeature image analysis. Using the manual digitized shorelines as the reference data, an average error of less than one pixel with standard deviation of 0.1320 can be obtained. These results prove that the proposed method is feasible for the identification and monitoring of thermokarst lakes on the QTP.

A Novel Adaptive Retinex for Image Enhancement

A novel adaptive Retinex algorithm for image enhancement is proposed in this paper. This method can automatically get the histogram truncation value on the basis of the actual pixels distribution of image. In experiment, we take multi images to test and verity the novel method and other methods. The experimental results show that the proposed method is superior to other methods.

Desertification Monitoring Using Remote Sensing and GIS Techniques

Proliferated in recent years the desertification phenomenon, and the desert areas started expanding at the expense of green areas, which affected the environment. This research focused on studying the desertification and its changes regarding to the time, through using different multi band satellite images for the area of interest in different times and studying the changes appear to the land cover and calculating the areas of each parameter to make the comparison for each environmental parameter (soil, agriculture, and water).

Reliable facial expression recognition for multi-scale images using weber local binary image based cosine transform features

Accurate recognition of facial expression is a challenging problem especially from multi-scale and multi orientation face images. In this article, we propose a novel technique called Weber Local Binary Image Cosine Transform (WLBI-CT). WLBI-CT extracts and integrates the frequency components of images obtained through Weber local descriptor and local binary descriptor. These frequency components help in accurate classification of various facial expressions in the challenging domain of multi-scale and multi-orientation facial images. Identification of significant feature set plays a vital role in the success of any facial expression recognition system. Effect of multiple feature sets with varying block sizes has been investigated using different multi-scale images taken from well-known JAFEE, MMI and CK+ datasets. Extensive experimentation has been performed to demonstrate that the proposed technique outperforms the contemporary techniques in terms of recognition rate and computational time.

고해상도 범색 영상을 위한 다중 단계 영상 복원

고해상도 범색 영상을 위한 다중 단계 영상 복원

Image Fusion using Variational Mode Decomposition

Background/Objectives: This paper introduced an image fusion algorithm based on Variational Mode Decomposition (VMD). Methods/Statistical Analysis: Image fusion is one of the image enhancement methods which results the image with better quality derived from a set of degraded images. Fused image contains more information than input images and it is efficient for visual perception and computer vision applications. This paper proposed an image fusion technique based on VMD for multi focus images. VMD has been a recently introduced non-recursive decomposition method, which decomposes the image into separate spectral bands called Intrinsic Mode Function (IMF) or modes. The modes are generated with respect to the associated central frequencies and they are band limited. Findings: A fusion rule based on weighing scheme is performed at the decomposition level for increasing the features by decreasing the mutual information. The reconstruction of the IMFs results the final fused image. The performance analysis of the proposed fusion method is experimented using standard objective quality metrics. The efficiency of the proposed method is determined by comparing the method with some state of the art methods. Application/Improvements: The image fusion using VMD is applicable to multi-resolution, multi model multi-sensor images.

Towards multi-stage texture-based active contour image segmentation

In this paper, we present a three-stage approach to incorporation of texture analysis into a two-dimensional active contour segmentation framework. This approach allows to utilise texture information alongside other image features. The proposed method starts with an initial unsupervised feature computation and selection, then moves to a fast contour evolution process and ends with a final refinement stage. The algorithm is designed to be general in its nature and not restricted to any particular texture feature extraction method. In this paper, the initial stage generates a set of feature maps consisting of grey-level co-occurrence matrix and Gabor features. The implementation makes an extensive use of hardware acceleration for efficient calculation of a relatively large number of features. The performance of the method was tested on various synthetic and natural images and compared with results of other algorithms.

Optimal Image Fusion

In this paper a saras-type image-fusion technique using hybrid SPIHT and SOMA. This method shows fusion of images and can be used for fusion of multi model image. It is concluded that fusion with advanced single levels offers better fusion quality. This method provides a comparative study between proposed & literature techniques and validation of the projected algorithm as Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE).

Study on optimization of a fewest layer measuring method of osteonecrotic area in femoral head

To optimize a measuring method of osteonecrotic area by analyzing the average osteonecrotic areas and osteonecritic volume using multi layer MR images. The MRI images of 87 cases (120 hips) of ONFH(ARCO II) were collected retrospectively from January 2011 to January 2012 in Wangjing Hospital of China Academy of Chinese Medicine Science. PHILIPS Achieval 1.5T MR was used to obtain coronal TSE T1 weighted (T1W) images. The scanning parameters were shown as follows:slice thickness, 3.5 mm; gap, 0.3 mm; images repetition time(TR), 500 ms;echo time(TE), 20 ms;field of view (FOV), 374 mm;total 12 layers. According to the distribution rule of osteonecrotic lesion, the layer of coronal T1-weighted imaging showing most of femoral neck was marked as layer 0(L0). The layers before L0 were marked in sequence L1, L2, L3, L4... , and the layers after L0 were marked in sequence L-1, L-2, L-3, L-4... . Auto CAD 2007 was used to measure the percentage of osteonecrotic area and calcu late the average data, and then decreased the layer from low to high layer successively based on frequency of osteonecrotic occurrences. First, the layer with lowest frequency of osteonecrotic occurrenoses L3 was removed, then calculated the average osteonecrotic area of the ramaining 8 layers. L5, L4, L-2, L-1, L3 layers were gradually removed, resulting in the calculation of avereage osteonecrotic areas in 7, 6, 5, 4, 3 layers. These areas were compared to the osteonecrotic volume in MR imagings, leading to the optimization of the fewest layer measuring method of osteonecrotic area using a statistical analysis. The percentage of osteonecrotic volume in 120 hips was 0.333±0.151. The average osteonecrotic areas of 9 to 3 layers were 0.321±0.117, 0.317±0.136, 0.312±0.147, 0.333±0.153, 0.348±0.172, 0.365±0.174, 0.377±0.202 respectively. There were no statistical differences of the average osteonecrotic areas and osteonecrotic volume in 9 to 3 layers(P>0.05), but when the osteonecrotic layers were reduced to 3, there were statistical differences(P<0.05). Total 120 hips were grouped according to osteonecrotic volume based on ARCO staging criteria, among them, 12 hips were grade A, 43 were B, 65 were C. According to average osteonecrotic areas of 4 layers, 10 hips were A grade, 32 were B, 78 were C. There were no statistical differences between two methods(P>0.05). There was a high degree of concordance among two methods. The results of 4(L0, L1, L2, L3) layers measuring method and osteonecrotic volume measuring method are similar. The 4 layers measuring method is an accurate, convenient, valuable method measuring the esteonecrotic area with the fewest layers, which is worth to be popularized in clinical application.

3D Reconstruction of Remote Sensing Image Using Region Growing Combining with CMVS-PMVS

The characteristics of remote sensing image is not obvious and can not reflect the be reconstructed the detail characters of the objects. For the sparse points in multi images fusion results, texture selection and it is relatively difficult problems, for accurate reconstruction of remote sensing image details, this paper presents CMVS-PMVS(cluster multi view stereo- patch based multi view stereo) combining with region growing for computing dense matching points. First, select some seed points by region growing algorithm, and find the matching relationship between the seed points, followed by the matching relation from the seed point to spread until to the entire image. Then the image set are clustered by CMVS in order to reduce the amount of data in the process of reconstruction, and the operation rate and reconstruction accuracy can be improved. PMVS reconstruction method is used to complete the reconstruction task by matching, expanding and filtering three steps. The experimental results showed that the point cloud is dense enough which are reconstructed by the 3d reconstruction algorithm based on regional growth combining CMVS-PMVS and well expressed the practical model of object reconstruction, the reconstruction of objects in remote sensing images has very strong practicability.

INVESTIGATIONS ON THE POSSIBILITIES OF MONITORING COASTAL CHANGES INCLUDING SHALLOW UNDER WATER AREAS WITH UAS PHOTO BATHMETRY

UAS become a very valuable tool for coastal morphology. Not only for mapping but also for change detection and a better understanding of processes along and across the shore. This contribution investigates the possibilities of UAS to determine the water depth in clear shallow waters by means of the so called "photo bathymetry". From the results of several test flights it became clear that three factors influence the ability and the accuracy of bathymetric sea floor measurements. Firstly, weather conditions. Sunny weather is not always good. Due to the high image resolution the sunlight gets focussed even in very small waves causing moving patterns on shallow grounds with high reflection properties, such as sand. This effect invisible under overcast weather conditions. Waves, may also introduce problems and mismatches. Secondly the quality and the accuracy of the georeferencing with SFM algorithms. As multi image key point matching will not work over water, the proposed approach will only work for projects closely to the coastline with enough control on the land. Thirdly the software used and the intensity of post processing and filtering. Refraction correction and the final interpolation of the point cloud into a DTM are the last steps. If everything is done appropriately, accuracies in the bathymetry in the range of 10 – 50 cm, depending on the water depth are possible.

INVESTIGATIONS ON THE POSSIBILITIES OF MONITORING COASTAL CHANGES INCLUDING SHALLOW UNDER WATER AREAS WITH UAS PHOTO BATHMETRY

Abstract. UAS become a very valuable tool for coastal morphology. Not only for mapping but also for change detection and a better understanding of processes along and across the shore. This contribution investigates the possibilities of UAS to determine the water depth in clear shallow waters by means of the so called "photo bathymetry". From the results of several test flights it became clear that three factors influence the ability and the accuracy of bathymetric sea floor measurements. Firstly, weather conditions. Sunny weather is not always good. Due to the high image resolution the sunlight gets focussed even in very small waves causing moving patterns on shallow grounds with high reflection properties, such as sand. This effect invisible under overcast weather conditions. Waves, may also introduce problems and mismatches. Secondly the quality and the accuracy of the georeferencing with SFM algorithms. As multi image key point matching will not work over water, the proposed approach will only work for projects closely to the coastline with enough control on the land. Thirdly the software used and the intensity of post processing and filtering. Refraction correction and the final interpolation of the point cloud into a DTM are the last steps. If everything is done appropriately, accuracies in the bathymetry in the range of 10 – 50 cm, depending on the water depth are possible.