Multiscale Local Feature Fusion: Marine Microalgae Classification for Few-Shot Learning

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.

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.

Differential physiological response of marine and freshwater microalgae to polystyrene microplastics

The appearance of pedestrians can vary greatly from image to image, and different pedestrians may look similar in a given image. Such similarities and variabilities in the appearance and clothing of individuals make the task of pedestrian re-identification very challenging. Here, a pedestrian re-identification method based on the fusion of local features and gait energy image (GEI) features is proposed. In this method, the human body is divided into four regions according to joint points. The color and texture of each region of the human body are extracted as local features, and GEI features of the pedestrian gait are also obtained. These features are then fused with the local and GEI features of the person. Independent distance measure learning using the cross-view quadratic discriminant analysis (XQDA) method is used to obtain the similarity of the metric function of the image pairs, and the final similarity is acquired by weight matching. Evaluation of experimental results by cumulative matching characteristic (CMC) curves reveals that, after fusion of local and GEI features, the pedestrian reidentification effect is improved compared with existing methods and is notably better than the recognition rate of pedestrian re-identification with a single feature.

Semi-supervised process fault classification based on convolutional ladder network with local and global feature fusion

In order to retrieve the image quickly and accurately in the massive image library, this paper proposes a fast retrieval method for massive face images based on global feature and local feature. Firstly, we use the local binary model feature (LBP) to extract the local face features, such as the tip of the nose, the mouth, and eye pupil etc. Then, the image global features are extracted and are integrated with the local features as our retrieval features. The principal component analysis (PCA) is used to reduce the dimensions of the features to 64, and the reduced dimension is encoded to generate an image signature, whose inverted index table is constructed for the image library and used for efficient retrieval. By testing on the 110,000 experimental datasets, the method can accurately retrieve the desired image within 0.3s using single-thread program.

Toxicity of metals to aquatic organisms is dependent on both external factors, such as exposure concentration and water quality parameters, and intracellular processes including specific metal-binding sites and detoxification. Current models used to predict copper toxicity in microalgae do not adequately consider these intracellular processes. This study compared the copper-binding proteins from four species of marine microalgae, Dunaliella tertiolecta, Tetraselmis sp., Phaedactylum tricornutum and Ceratoneis closterium, in controls (no added copper) and following a 72-h exposure to copper (sufficient to inhibit growth by approximately 50%). Cells were lysed by sonication, which was optimised to obtain 54-94% cell rupture for the different algae. Cell lysates were processed by immobilised metal affinity chromatography (IMAC) using Cu(2+) as the bound metal (i.e. Cu-IMAC). Bound proteins were subsequently analysed by SDS-PAGE, comparing proteins recovered from algae that were exposed to copper versus untreated control cells. Individual proteins for which copper exposure resulted in changes to proteins present were excised from gels and further analysed by nano LC ESI-MS/MS; proteins were identified using the Mascot database. Proteins identified in this way included heat-shock proteins, rubisco, α- and β-tubulins and ATP synthase (β subunit). The results established that Cu-IMAC is a useful approach to identify proteins involved in copper binding in algae. This study identified several proteins that may play an active role in responses to copper toxicity in marine microalgae.

Human activity recognition is an important and challenging topic for computer vision research community. Action representation using deep learning models are currently the dominant technique compared with other methods. However, supervised convolutional neural networks require large computational and memory resources to optimize their parameters. Recently, a simple unsupervised deep learning architecture Principal Component Analysis Network (PCANet) has emerged as an alternative of Convolutional Neural Networks (CNNs) and has significant accomplishments in various vision applications. Meanwhile, encoding and representation techniques using Bag of Words (BoW) and Vector of Locally Aggregated Descriptors (VLAD) have demonstrated great success for several visual tasks specifically in activity recognition. This work presents a novel human activity recognition technique by combining global and local features of PCANet with BoW and VLAD encoding schemes. Both global and local features are learned by PCANet utilizing selected frames from each action video. After that the dimensionality of these features is decreased via Whitening PCA (WPCA). Then encoding schemes are applied on both features to represent the final descriptors for each action. Ultimately, Support Vector Machines classifier (SVM) is trained for recognition process. Several experiments are conducted on UCF sports dataset to evaluate our method. All experimental results utilizing leave-one-out-cross validation (LOOCV) strategy are satisfactorv and comparable.

Vision graph convolutional network for underwater image enhancement

A wavelet-based local and global feature fusion network (LAGN) is proposed for low-light image enhancement, aiming to enhance image details and restore colors in dark areas. This study focuses on addressing three key issues in low-light image enhancement: Enhancing low-light images using LAGN to preserve image details and colors; extracting image edge information via wavelet transform to enhance image details; and extracting local and global features of images through convolutional neural networks and Transformer to improve image contrast. Comparisons with state-of-the-art methods on two datasets verify that LAGN achieves the best performance in terms of details, brightness, and contrast.

Accurate and interpretable AI models play a critical role in medical image analysis. However, despite advancements in explainable AI (XAI), existing methods struggle with inconsistent interpretability. To overcome this limitation, we introduce the Integrated Latent and Attention Mapping (ILAM) framework, which enhances both classification accuracy and explainability by fusing local and global feature representations. ILAM integrates a custom-designed Autoencoder (AE) with a Vision Transformer (ViT), where the AE learns fine-grained local features through unsupervised patchwise image reconstruction in the latent space. These local features are then fused with global representations extracted by ViT, creating a hybrid model that improves both performance and post hoc interpretability. To refine explainability, ILAM incorporates a modified attention rollout mechanism, which recursively aggregates latent feature representations and attention weights to produce precise and stable activation maps. We evaluate ILAM on three publicly available medical imaging datasets-BreakHis, Chest X-Ray, and Retinal, demonstrating its superior performance over transformer-based models such as ViT, DeiT, CvT, and SwinT. ILAM consistently generates detailed and reliable activation maps, providing clearer visualizations of critical image regions influencing model decisions. By effectively combining local and global feature fusion, ILAM establishes itself as a robust and interpretable framework for medical image classification.

In many real-world problems, collecting a large number of labeled samples is infeasible. Few-shot learning (FSL) is the dominant approach to address this issue, where the objective is to quickly adapt to novel categories in presence of a limited number of samples. FSL tasks have been predominantly solved by leveraging the ideas from gradient-based meta-learning and metric learning approaches. However, recent works have demonstrated the significance of powerful feature representations with a simple embedding network that can outperform existing sophisticated FSL algorithms. In this work, we build on this insight and propose a novel training mechanism that simultaneously enforces equivariance and invariance to a general set of geometric transformations. Equivariance or invariance has been employed standalone in the previous works; however, to the best of our knowledge, they have not been used jointly. Simultaneous optimization for both of these contrasting objectives allows the model to jointly learn features that are not only independent of the input transformation but also the features that encode the structure of geometric transformations. These complementary sets of features help generalize well to novel classes with only a few data samples. We achieve additional improvements by incorporating a novel self-supervised distillation objective. Our extensive experimentation shows that even without knowledge distillation our proposed method can outperform current state-of-the-art FSL methods on five popular benchmark datasets.

Approximately 48% of Americans use prescription drugs within each 30-day period, and there are signs this trend is increasing. Although many studies track pharmaceuticals’ fates in contaminating waterways, only fairly recent efforts have examined the potential impacts of these drugs on non-target organisms. The antibiotics sulfamethoxazole and trimethoprim, often prescribed together to treat bacterial infections, have been detected worldwide in marine and estuarine environments at concentrations up to 700-800 ng/L each. Toxic effects of these drugs have been identified in freshwater organisms, with synergistic effects observed in short-term studies of mixtures of the two; however, little research has examined possible sub-lethal and longer-term effects of antibiotics in environmentally-relevant mixture concentrations on marine organisms. I examined the long-term effects of mixtures of these two antibiotics in species of a marine system: marine microalgal species, and marine mussels, to determine whether the levels currently present in waterways affect organism health and reproduction. Microalgal species may be among the most sensitive organisms to pharmaceutical contaminants based on ecotoxicity research. I exposed three species of marine microalgae (Isochrysis galbana, Chaetoceros neogracile, and Nannochloropsis oculata) to environmentally-relevant mixtures of sulfamethoxazole and trimethoprim and examined their three-week growth rates. I found that for each species, the antibiotic treatments significantly suppressed algal growth. Specifically, I found that sulfamethoxazole was a driving factor in suppression of C. neogracile and I. galbana growth, while N. oculata responded more sensitively to a broader range of treatment mixture levels, which also included trimethoprim-only treatment groups and mixtures. These results on marine microalgae address critical data gaps, and identify the impacts of pharmaceuticals on marine primary producers, which could have direct ecosystem implications to higher trophic levels. Antibiotic pharmaceuticals can also affect marine mussel health, based on previous study of sub-cellular endpoints. I hypothesized that the important benthic foundational species, the Mytilus californianus mussel, would be significantly impacted by long-term 12-week exposure to environmentally-relevant concentrations and mixtures of sulfamethoxazole and trimethoprim. Specifically, I measured growth rate, feeding rate, condition index, and gonando-somatic index as indicators of whole-organism and reproductive health. Sulfamethoxazole concentrations, in particular, and trimethoprim to a lesser extent, suppressed mussel growth and significantly affected condition index and gonandosomatic index over time. The results of this study offer an understanding of how an intertidal system responds to chronic presence of antibiotic mixtures in the water, and a more complete picture of the environmental consequences of pharmaceutical contaminants released into marine ecosystems

Natural chemicals produced by marine microalgae as predator deterrents can be used to control ciliates contamination in microalgal cultures

A review was made of the literature concerning the metabolism of fatty acids in the dairy cow, and the potential of increasing the content of the essential fatty acids eicosapentaenoic acid (EPA; C20:5) and docosahexaenoic acid (DHA; C22:6) in cows' milk. EPA and DHA are two n-3 polyunsaturated fatty acids, which can be synthesised from a-linolenic acid (LNA, C18:3). Both EPA and DHA have anti-thrombotic and anti-carcinogenic effects. The UK Department of Health has recommended that the daily human intake of total n-3 PUFA should be 0.2 g/d, which is approximately double the current average intake. The richest source of dietary EPA and DHA is marine products, but the consumption of fish in the UK is declining. Milk fat consumption constitutes 30% of total fat consumption in the UK diet, and so enhancing the concentration of EPA and DHA in milk fat could be an effective means of increasing the consumption of these fatty acids in the UK. Grass and linseed are rich sources of LNA for dairy cows, while fish oil or marine algae are the richest sources of EPA and DHA. However, a large proportion of these fatty acids ( ca 0.9) are hydrogenated to more saturated fatty acids by the microorganisms in the rumen. The proportion of absorbed n-3 PUFA is therefore only about 7 g/kg ingested n-3 PUFA. Absorbed EPA and DHA are transported around the body as phospholipids, associated with the high-density lipoprotein fraction of the plasma. This fraction is not a good substrate for lipoprotein lipase, which is the enzyme in the mammary gland that removes fatty acids from the circulation for incorporation into milk fat. The uptake by the mammary gland of absorbed EPA and DHA is therefore extremely low. The concentration of EPA and DHA in milk fat therefore remains low even when the diet is enriched with these fatty acids. It is possible that the EPA and DHA present in the milk arise from de novo synthesis of these acids from LNA in the mammary gland. However, while EPA is synthesised from LNA, there appears to be a block on the subsequent synthesis of DHA from EPA. EPA and DHA are not stored in adipose tissue triacylglycerols, and they are not used for oxidative metabolism. There is therefore little competition for circulating EPA and DHA between the mammary gland and adipose tissue, or between secretion and oxidation. There is some potential storage of these acids in the phospholipids of muscle tissue and adipose tissue. Grass is a rich source of LNA, and the LNA content of milk can be enhanced by increasing the herbage content of the diet. This potentially provides more precursors of EPA and DHA. However, increasing the proportion of forage in the diet also tends to increase the extent of rumen biohydrogenation of PUFA as this process is encouraged by higher rumen pH and reduced propionate production. Including either fish oil or some species of marine microalgae in the diet increases the EPA and DHA content of milk. However, these oils can adversely affect rumen function, so that total dry matter intake and milk fat content are reduced. Feeding fish oils to ruminant animals may also not be a sustainable practice in the longer-term. Protecting marine oils from rumen biohydrogenation should prevent the adverse effects on rumen function. However, the evaluation of a protected marine microalgae product suggested that the protection was far from complete, and a reliable means of treating algae to protect it from rumen metabolism still awaits development. Further enhancement of the EPA and DHA content of cows' milk requires more research. Issues, which need to be addressed, are the development of a suitable oil supplement, and a reliable means of protecting it from rumen biohydrogenation. A means of transporting the absorbed EPA and DHA in the circulation so that it is readily taken up by the mammary gland for incorporation into the milk fat also needs to be developed. Changes in the marketing of milk that has been improved in this way would also be required to meet the increased cost of producing this nutritionally enhanced milk.