Infrared point target detection based on multiscale homogeneous feature fusion

The fully convolutional networks (FCNs) have been widely applied in numerous medical image segmentation tasks. However, tissue regions usually have large variations of shape and scale, so the ability of neural networks to learn multi-scale features is important to the segmentation performance. In this paper, we improve the network for multi-scale feature fusion, in the medical image segmentation by introducing two feature fusion modules: i) global attention multi-scale feature fusion module (GMF); ii) local dense multi-scale feature fusion module (LMF). GMF aims to use global context information to guide the recalibration of low-level features from both spatial and channel aspects, so as to enhance the utilization of effective multi-scale features and suppress the noise of low-level features. LMF adopts bottom-up top-down structure to capture context information, to generate semantic features, and to fuse feature information at different scales. LMF can integrate local dense multi-scale context features layer by layer in the network, thus improving the ability of network to encode interdependent relationships among boundary pixels. Based on the above two modules, we propose a novel medical image segmentation framework (GLF-Net). We evaluated the proposed network and modules on challenging brain tumor segmentation and pancreas segmentation datasets, and very competitive performance has been achieved.

In order to achieve a single-frame enhancement and detection of dim point targets efficiently, a new method based on morphological top-hat transformation and modified non-local means (NL-means, NLM) for target enhancement is presented. After enhancing dim point targets, an estimating algorithm called local reverse entropy is applied to get candidate targets, and obtains detecting results in the end. In this model, white top-hat and black top-hat are combined to obtain pre-processed image, then the targets are enhanced for the second time through modified NL-means algorithm. The background in the residual image of enhanced image and original image is mostly suppressed, then it is regarded as the input of local reverse entropy estimation method (LREM). Detecting results can be obtained by setting proper thresholds. The 4 × 4 weak lattice targets with different brightness are superimposed on different infrared image backgrounds. The experimental results show that, when the SCR is low (SCR≈1), the detection algorithm model proposed in this paper has higher SCR gain than other target-enhancing algorithms such as TDLMS, max-median, max-mean, non-local means, etc. and the detection performance is the best.

This paper proposes a hybrid method for face recognition using local features and statistical feature extraction methods. First, a dense set of local feature points are extracted in order to represent a facial image. Each local feature point is described by the keypoint descriptor defined by SIFT feature. Then, the statistical feature extraction methods, PCA and LDA, are applied to the set of local feature descriptors in order to find low dimensional features. With the obtained low dimensional feature vectors, we can conduct face recognition task efficiently using a simple classifier. Through computational experiments on benchmark data sets, we show that the proposed method is superior to the conventional PCA and LDA in the classification performance. In addition, we also show that the proposed method can achieve remarkable improvement in the processing time compared to the conventional keypoint matching methods proposed for local features.KeywordsFace RecognitionLocal FeatureSift FeatureGray Level IntensityConventional Statistical MethodThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

In colonoscopy examinations, the dark imaging environment of the intestine may result in low-quality scene acquisition and pose challenges for subsequent disease diagnosis. Despite the development of numerous low-light image enhancement (LIE) methods to improve the quality of enhanced colonoscopy images (ECIs), various artifacts such as abnormal brightness and low contrast are still induced. To fairly compare various LIE algorithms and accurately measure the visual quality of ECIs, we develop a novel blind quality evaluator for ECIs by integrating local and global statistical features. For global features, we first extract multiscale gradient similarity features in the multiscale space domain and multi-resolution information entropy features in the discrete cosine transform (DCT) domain to capture their sharpness degradation. Second, we utilize asymmetric generalized Gaussian distribution (AGGD) models to extract fitting parameters in color-opponent components to measure color distortion in ECIs. In addition, a set of intermediate ECIs are generated by adjusting the brightness information, from which perceptual features are computed to capture global brightness variations. For local features, we utilize local binary pattern (LBP) descriptors in the gradient domain and combine them with contrast energy features and local color-aware features to capture local quality degradation. Finally, by integrating both local and global perceptual features, a regression function is employed to build a connection between quality-aware features and subjective ratings. Extensive experiments conducted on publicly available ECIQAD dataset demonstrate that our method outperforms current mainstream blind quality metrics.

Image enhancement plays a crucial role in computer vision by improving visual quality while minimizing distortion. Traditional methods enhance images through pixel value transformations, yet they often introduce new distortions. Recent advancements in deep learning-based techniques promise better results but challenge the preservation of image fidelity. Therefore, it is essential to evaluate the visual quality of enhanced images. However, existing quality assessment methods frequently encounter difficulties due to the unique distortions introduced by these enhancements, thereby restricting their effectiveness. To address these challenges, this paper proposes a novel blind image quality assessment (BIQA) method for enhanced natural images, termed multi-scale local feature fusion and global feature representation-based quality assessment (MLGQA). This model integrates three key components: a multi-scale Feature Attention Mechanism (FAM) for local feature extraction, a Local Feature Fusion (LFF) module for cross-scale feature synthesis, and a Global Feature Representation (GFR) module using Vision Transformers to capture global perceptual attributes. This synergistic framework effectively captures both fine-grained local distortions and broader global features that collectively define the visual quality of enhanced images. Furthermore, in the absence of a dedicated benchmark for enhanced natural images, we design the Natural Image Enhancement Database (NIED), a large-scale dataset consisting of 8,581 original images and 102,972 enhanced natural images generated through a wide array of traditional and deep learning-based enhancement techniques. Extensive experiments on NIED demonstrate that the proposed MLGQA model significantly outperforms current state-of-the-art BIQA methods in terms of both prediction accuracy and robustness.

A multi-scale semantic feature fusion method for remote sensing crop classification

Infrared (IR) point target detection and tracking is challenging due to lack of texture and detailed information of small dim targets. The problem becomes even more challenging when there is a real-time requirement too. A key goal in robust detection is to reduce missed detections (MD) and false alarms (FA). Complex traditional state-of-the-art point target detection algorithms may give accurate results but typically incur larger execution times which renders them unsuitable for real-time applications. On the other hand, deep learning methods for point target detection do not generalize satisfactorily with domain shifts. In this work, we address the problem of real-time detection and tracking of multiple point targets in diverse background conditions. We apply a ’top-hat’ operator as the first stage of our detection algorithm and this is followed by a systematic thresholding scheme to yield a good balance between MD and FA. This is followed by a methodology to find the exact target position from the detections, and a track association scheme in conjunction with Kalman filter for state estimation. Based on the proposed approach for target detection and track association, we perform 2D tracking of image coordinates as well as 3D tracking of azimuth and elevation angles. We verify the effectiveness of our method on 12 different IR image sequences over existing state-of-the-art methods in terms of accuracy as well as speed.

Pose-invariant face recognition using facial landmarks and Weber local descriptor

The classification of high-resolution (HR) synthetic aperture radar (SAR) images is of great importance for SAR scene interpretation and application. However, the presence of intricate spatial structural patterns and complex statistical nature makes SAR image classification a challenging task, especially in the case of limited labeled SAR data. This paper proposes a novel HR SAR image classification method, using a multi-scale deep feature fusion network and covariance pooling manifold network (MFFN-CPMN). MFFN-CPMN combines the advantages of local spatial features and global statistical properties and considers the multi-feature information fusion of SAR images in representation learning. First, we propose a Gabor-filtering-based multi-scale feature fusion network (MFFN) to capture the spatial pattern and get the discriminative features of SAR images. The MFFN belongs to a deep convolutional neural network (CNN). To make full use of a large amount of unlabeled data, the weights of each layer of MFFN are optimized by unsupervised denoising dual-sparse encoder. Moreover, the feature fusion strategy in MFFN can effectively exploit the complementary information between different levels and different scales. Second, we utilize a covariance pooling manifold network to extract further the global second-order statistics of SAR images over the fusional feature maps. Finally, the obtained covariance descriptor is more distinct for various land covers. Experimental results on four HR SAR images demonstrate the effectiveness of the proposed method and achieve promising results over other related algorithms.

Independence of Synthetic Aperture Radar (SAR) images from weather and sun illumination helps us in maritime surveillance like an oil spill, to identify the illegal fishery activity and unauthorized marine vessels. Detection of marine vessels is like a point target detection in coarse to moderate resolution images. Detection of point targets is adversely affects by speckle noise. Object detection in SAR images, without explicitly reducing the speckle noise is one of the challenging task. In this paper two algorithms are presented, which show how improvement in the power of point target lead us to reduced number of false alarms. The first algorithm has three parts. First part uses Grave matrix, which is a ( $2\times 2$ ) Hermitian power matrix, generated by the multiplication of Sinclair matrix and conjugate of Sinclair matrix. Second part uses discrimination criteria to discriminate between clutter and vessels based on eigenvalue of the Grave matrix (G). The third part, fill the gaps by using morphological dilation. The only difference between the first and the second algorithms is that, in the second algorithm we uses Sinclair matrix (S2) instead of Grave matrix. Both the algorithms tested on two full-polarization (HH, HV, VH, VV) datasets and the results show the importance of scattering power matrix (G) as compared to scattering matrix (S2) for point target detection. The first data set is of size $498^{\ast}498$ captured by AlOS-1 PALSAR L-band data, covering the coastal region of Singapore. The second data set is of size $472^{\ast}472$ , covering the coast of Vancouver acquired in C-band by Radarsat-2.

Non-uniformity correction of infrared focal plane array in point target surveillance systems

In point target optical detection, the selection of the detection spectrum largely determines the upper limit of detection capability. In this paper, we analyze the main factors affecting point target detection and propose a spectral segment selection method for point target detection based on constant false alarm. The method uses the probability of detection as an evaluation index and considers the target and background radiation properties, atmospheric radiative transfer, background clutter and sensor noise levels. The method can calculate the optimal spectral band for its detection for different cases of target and background information. In this paper, we analyze the clutter levels of different surface types of background using MODIS data, and take typical airborne target detection as an example. We calculate the range of spectral bands whose detection probabilities meet the requirements under several conditions to verify its feasibility. The results show that the point target detection spectra satisfying the false alarm rate and detection probability requirements can be obtained by the method in this paper

In medical imaging diagnosis, accurate segmentation of the knee joint can help doctors better observe and diagnose lesions, thereby improving diagnostic accuracy and treatment effectiveness. Vision Mamba mainly relies on the State Space Model (SSM) for feature modeling, which excels at capturing global contextual information but cannot capture local texture features. Moreover, features of different scales are not effectively integrated, resulting in the model's weak segmentation ability on small-scale tissues (such as cartilage areas). To this end, this study proposed a novel multi-scale Vision Mamba Unet (VM-Unet) framework named MSPF-VM-Unet to perform the segmentation on the femur, tibia, femoral cartilage, and tibial cartilage in knee MRI images. The proposed MSPF-VM-Unet extends VM-Unet by introducing a designed multi-scale pyramid feature extraction network named MPSK, which synergizes multi-resolution feature extraction with channel-space attention. MPSK network enhances multi-scale local feature extraction through Selective Kernel (SK) convolution and pyramid pooling. The network merges the overall context information extracted by the Vision Mamba encoder to achieve the coordinated optimization of a multi-scale hierarchical feature fusion mechanism and global long-range dependency modeling. The results of the comparative experiments on the OAI-ZIB dataset indicate that MSPF-VM-Unet significantly improves the boundary accuracy and regional consistency of the MRI tibiofemoral joint tissue structure.

This paper deals with the detection of dim point targets in infrared images. Dim point targets detection is always a difficulty in information processing. Researchers have proposed many effective methods; this paper introduces a new method. Whereas difference method has obtained good result in one dimensional signal processing, this paper manages to apply it to two dimensional signal processing, that is to say, dim point targets detection in infrared images of low SNR. The image background is color noise, and its column correlation is strong. So, ordinary methods probably lose the true targets because of the strong color noise, but unlike them, difference method can overcome this shortcoming, it can eliminate correlation noise and enhance useful information, finally pick out the probable targets from noise background. In the paper, the method was given a more extensive account. In order to improve the detection effect, we utilize prefilter. The prefilter is realized by the alpha filter. Because the same targets have same location in more than three frames, using alpha filter can utilize the information of adjacent frames, increase the SNR of the raw data and reduce the noise of the images. >

Although Transformer-based methods have achieved remarkable performance in the field of long-term series forecasting, they can be computationally expensive and lack the ability to specifically model local features as CNNs. CNN-based methods, such as temporal convolutional network (TCN), utilize convolutional filters to capture local temporal features. However, the intermediate layers of TCN suffer from a limited effective receptive field, which can result in the loss of temporal relations during global feature extraction.To solve the above problems, we propose to combine local features and global correlations to capture the overall view of time series (e.g. fluctuations, trends). To fully exploit the underlying information in the time series, a multi-scale branch structure is adopted to model different potential patterns separately. Each pattern is extracted using a combination of interactive learning convolution and causal frequency enhancement to capture both local features and global correlations. Furthermore, our proposed method,multi-scale local-global feature learning network (MLGN), achieves a time and memory complexity of O(L) and consistently achieve state-of-the-art results on six benchmark datasets. In comparision with previous best method Fedformer, MLGN yields 12.98% and 11.38% relative improvements for multivariate and univariate time series, respectively. Our code and data are available on Github at https://github.com/Zero-coder/MLGN.