Low Light Conditions Research Articles

Tightly coupled visual-inertial fusion with image enhancement for robust positioning

Abstract Traditional vision-based inertial odometry (VIO) suffers from significant visual degradation, which substantially impacts state estimation in challenging lighting environments. Thermal imaging cameras capture images based on the thermal radiation of objects, rendering them impervious to lighting variations. However, integrating thermal infrared data into conventional visual odometry poses challenges due to its low texture, poor contrast, and high noise levels. In this paper, we propose a tightly coupled approach that seamlessly integrates information from visible light cameras, thermal imaging cameras, and inertial measurement units (IMUs). First, we employ adaptive bilateral filtering and Sobel gradient enhancement to smooth infrared images, thereby reducing noise and enhancing edge contrast. Second, we leverage the Sage-Husa adaptive filter in conjunction with iterative Kalman filtering (IEKF) to effectively mitigate the impact of non-Gaussian noise on the system. Finally, we conduct comprehensive evaluations of the proposed system using both open datasets and real-world experiments across four distinct scenarios: normal lighting, low-light conditions, low-light conditions with camera shake, and challenging lighting environments. Comparative analysis reveals that our method outperforms IEKF, yielding a reduction in localization error measured by root mean square error (RMSE) by 58.69%, 57.24%, 60.23%, and 30.87% in these respective scenarios.

LIASM-NRID: Constructing an atmospheric scattering model for low-light conditions and dehazing nighttime road images

Dehazing of nighttime road images holds significant practical relevance for autonomous driving and intelligent transportation systems operating in nocturnal conditions. To address the limitations of traditional atmospheric scattering models in accurately describing the imaging process in low illumination environments such as nighttime, this study constructs a low illumination environment atmospheric scattering model. Based on this newly constructed model, we propose an algorithm specifically designed for dehazing road images at nighttime. The proposed model incorporates an incident light attenuation factor, an additive noise factor, and a global compensation factor related to the scene's depth. The proposed nighttime road image dehazing algorithm firstly suppresses the interference of additive noise in the original image using wavelet decomposition and soft thresholding operation. Then, a transmittance-solving method containing the primary structural information of the image is designed based on the color attenuation linear model, and the solved transmittance is refined based on the interval gradient image texture structure method. Consequently, the global light map estimation process is formulated as an optimization problem to ensure the accuracy of the atmospheric light value calculation. In the final stages, a novel RGB color equalization method is introduced to address the issue of non-uniform incident light color bias in the dehazed images. Additionally, to eliminate the halo problem in the dehazed images, the adaptive histogram equalization algorithm is modified by incorporating a normalized gamma correction function. The experimental results show that the proposed nighttime road image dehazing algorithm exhibits robust performance and generality across various nighttime scenes. Compared with the representative algorithms for nighttime image dehazing, it can get more excellent dehazing results and has more outstanding performance in all numerical evaluation indexes.

Low-light wheat image enhancement using an explicit inter-channel sparse transformer

Low-light image enhancement poses a significant challenge in agricultural settings, particularly for time-series captured wheat images. The use of time-series captured image data can improve the accuracy of wheat yield prediction by analyzing the growth status and characteristics of wheat. Given the difficulty of capturing paired wheat images in various lighting conditions, a method based on the Global Wheat Head Detection (GWHD) dataset is presented to synthesize pairs of low-light/normal-light wheat images. This approach expands to create a new dataset, LN-GWHD, specifically for the task of enhancing wheat images in low-light conditions. Importantly, it enables a more extensive evaluation of wheat image restoration tasks and performance accuracy under more controlled and diverse conditions. It is worth noting that this method can be extended to a wide range of scenarios for the enhancement of low-light situations. Where low-light conditions are common due to various factors such as weather, existing methods based on the transformer architecture have achieved state-of-the-art (SOTA) enhancement performance by capturing relationships between low-light feature channels. However, these approaches perform the similarity computation on all tokens and may introduce redundant features, limiting their ability to focus on critical information for high-quality image reconstruction. To recover more high-quality wheat image information, the explicit inter-channel sparse transformer (EIST) network is proposed, which is specifically designed to recover wheat images under low-light conditions. EIST comprises multiple blocks, each featuring explicit sparse top-k attention (ESTA) and bilateral gated feed-forward network (BGFN). ESTA dynamically selects the most critical information by focusing on the top-k key labels in each query, while BGFN enhances valid information interaction through re-double filtering. The overall architecture employs transformers to learn correlations and features between different image channels, resulting in noise reduction, detail enhancement, and improved clarity. Additionally, the Fourier spectrum loss is introduced to constrain the image reconstruction in the frequency domain, allowing more detailed information to be recovered. Extensive experiments on the LN-GWHD datasets demonstrate that EIST surpasses SOTA methods. Furthermore, our approach achieves superior results in detecting low-light wheat ears during back-end evaluation.

Visible characteristics and durability assessment of methyl methacrylate-based luminescent road marking using unencapsulated SrAl2O4: Eu2+, Dy3+ under coupling service conditions

Luminescent road markings (LRMs) have attracted growing attention due to the enhanced road visibility in low-light conditions and energy-saving benefits. However, high-performance strontium aluminate phosphors (SrAl2O4: Eu2+, Dy3+) are prone to hydrolyze while the conventional encapsulation methods of phosphors produces luminance loss. This paper presented a novel perspective using methyl methacrylate (MMA)-based two-component resins with unencapsulated phosphor (UP) to prepare LRMs with balanced luminescent properties road durability. Luminescent properties, skid resistance, impact resistance, and abrasion rate were evaluated as the response value to optimize the formulations through design of experimental (DOE) methods. The results show that LRMs exhibit outstanding water resistance and abrasion resistance after soaking for 120 hours coupling with wet-wheel abrasion for 20 minutes due to effective bonding between the carboxylic acid-based resin and phosphors. The luminance loss is less than 5 % after undergoing three-wheel abrasion following 480 hours of ultraviolet aging. Furthermore, the macroscopic performance was integrated with microscopic properties through morphological, elemental, and functional group analysis. The results demonstrated that the addition of plasticizer balanced the relationship among the mechanical properties by transforming the curing structure of the coating from resin-encapsulated granular protrusions into layer-like structures.

Real-Time Attentive Dilated U-Net for Extremely Dark Image Enhancement

Images taken under low-light conditions suffer from poor visibility, color distortion, and graininess, all of which degrade the image quality and hamper the performance of downstream vision tasks, such as object detection and instance segmentation in the field of autonomous driving, making low-light enhancement an indispensable basic component of high-level visual tasks. Low-light enhancement aims to mitigate these issues, and has garnered extensive attention and research over several decades. The primary challenge in low-light image enhancement arises from the low signal-to-noise ratio caused by insufficient lighting. This challenge becomes even more pronounced in near-zero lux conditions, where noise overwhelms the available image information. Both traditional image signal processing pipeline and conventional low-light image enhancement methods struggle in such scenarios. Recently, deep neural networks have been used to address this challenge. These networks take unmodified RAW images as input and produce the enhanced sRGB images, forming a deep learning based image signal processing pipeline. However, most of these networks are computationally expensive and thus far from practical use. In this article, we propose a lightweight model called attentive dilated U-Net (ADU-Net) to tackle this issue. Our model incorporates several innovative designs, including an asymmetric U-shape architecture, dilated residual modules for feature extraction, and attentive fusion modules for feature fusion. The dilated residual modules provide strong representative capability, whereas the attentive fusion modules effectively leverage low-level texture information and high-level semantic information within the network. Both modules employ a lightweight design but offer significant performance gains. Extensive experiments demonstrate that our method is highly effective, achieving an excellent balance between image quality and computational complexity—that is, taking less than 4ms for a high-definition 4K image on a single GTX 1080Ti GPU and yet maintaining competitive visual quality. Furthermore, our method exhibits pleasing scalability and generalizability, highlighting its potential for widespread applicability.

Research on Unsupervised Low-Light Railway Fastener Image Enhancement Method Based on Contrastive Learning GAN

The railway fastener, as a crucial component of railway tracks, directly influences the safety and stability of a railway system. However, in practical operation, fasteners are often in low-light conditions, such as at nighttime or within tunnels, posing significant challenges to defect detection equipment and limiting its effectiveness in real-world scenarios. To address this issue, this study proposes an unsupervised low-light image enhancement algorithm, CES-GAN, which achieves the model’s generalization and adaptability under different environmental conditions. The CES-GAN network architecture adopts a U-Net model with five layers of downsampling and upsampling structures as the generator, incorporating both global and local discriminators to help the generator to preserve image details and textures during the reconstruction process, thus enhancing the realism and intricacy of the enhanced images. The combination of the feature-consistency loss, contrastive learning loss, and illumination loss functions in the generator structure, along with the discriminator loss function in the discriminator structure, collectively promotes the clarity, realism, and illumination consistency of the images, thereby improving the quality and usability of low-light images. Through the CES-GAN algorithm, this study provides reliable visual support for railway construction sites and ensures the stable operation and accurate operation of fastener identification equipment in complex environments.

Notes on the experimental laboratory and field culture of Lessonia variegata (Laminariales)

ABSTRACT The kelp Lessonia variegata J. Agardh (Lessoniaceae) is a potential aquaculture target for bioactive compounds. We investigated the control and management of Lessonia gametogenesis, the settlement and retention of sporophytes on various culture strings, and the upscaling of string settlement for field trials. In the laboratory we held gametophytes at low light conditions for 2.5 months before altering conditions to stimulate them into the next part of the life cycle. Settlement and development of sporophytes was highest on Korean culture string, followed by Builders Line, however after agitation, fewer sporophytes were dislodged from the Builders Line. Larger scale settlement of spores onto 15 m lengths of Builders Line was successful, with sporophytes reaching ∼5 mm length in 46 days. Eight weeks after out-planting, the sporophytes were heavily fouled with other algae. Although the results presented here were small-scale and lack replication, they add to the existing knowledge on the biology and cultivation of this species.

The Effects of Light Level and Signal-to-Noise Ratio on the Task-Evoked Pupil Response in a Speech-in-Noise Task.

There is increasing interest in the measurement of cognitive effort during listening tasks, for both research and clinical purposes. Quantification of task-evoked pupil responses (TEPRs) is a psychophysiological method that can be used to study cognitive effort. However, light level during cognitively demanding listening tasks may affect TEPRs, complicating interpretation of listening-related changes. The objective of this study was to examine the effects of light level on TEPRs during effortful listening across a range of signal-to-noise ratios (SNRs). Thirty-six adults without hearing loss were asked to repeat target sentences presented in background babble noise while their pupil diameter was recorded. Light level and SNRs were manipulated in a 4 × 4 repeated-measures design. Repeated-measures analyses of variance were used to measure the effects. Peak and mean dilation were typically larger in more adverse SNR conditions (except for SNR -6 dB) and smaller in higher light levels. Differences in mean and peak dilation between SNR conditions were larger in dim light than in brighter light. Brighter light conditions make TEPRs less sensitive to variations in listening effort across levels of SNR. Therefore, light level must be considered and reported in detail to ensure sensitivity of TEPRs and for comparisons of findings across different studies. It is recommended that TEPR testing be conducted in relatively low light conditions, considering both background illumination and screen luminance. https://doi.org/10.23641/asha.25676538.

Amphistomy: stomata patterning inferred from 13C content and leaf-side-specific deposition of epicuticular wax.

The benefits and costs of amphistomy (AS) vs. hypostomy (HS) are not fully understood. Here, we quantify benefits of access of CO2 through stomata on the upper (adaxial) leaf surface, using 13C abundance in the adaxial and abaxial epicuticular wax. Additionally, a relationship between the distribution of stomata and epicuticular wax on the opposite leaf sides is studied. We suggest that the 13C content of long-chain aliphatic compounds of cuticular wax records the leaf internal CO2 concentration in chloroplasts adjacent to the adaxial and abaxial epidermes. This unique property stems from: (1) wax synthesis being located exclusively in epidermal cells; and (2) ongoing wax renewal over the whole leaf lifespan. Compound-specific and bulk wax 13C abundance (δ) was related to amphistomy level (ASL; as a fraction of adaxial in all stomata) of four AS and five HS species grown under various levels of irradiance. The isotopic polarity of epicuticular wax, i.e. the difference in abaxial and adaxial δ (δab - δad), was used to calculate the leaf dorsiventral CO2 gradient. Leaf-side-specific epicuticular wax deposition (amphiwaxy level) was estimated and related to ASL. In HS species, the CO2 concentration in the adaxial epidermis was lower than in the abaxial one, independently of light conditions. In AS leaves grown in high-light and low-light conditions, the isotopic polarity and CO2 gradient varied in parallel with ASL. The AS leaves grown in high-light conditions increased ASL compared with low light, and δab - δad approached near-zero values. Changes in ASL occurred concomitantly with changes in amphiwaxy level. Leaf wax isotopic polarity is a newly identified leaf trait, distinguishing between hypo- and amphistomatous species and indicating that increased ASL in sun-exposed AS leaves reduces the CO2 gradient across the leaf mesophyll. Stomata and epicuticular wax deposition follow similar leaf-side patterning.

Motion segmentation with event camera: [formula omitted]-patches optical flow estimation and Pairwise Markov Random Fields

This paper proposes the N-patches optical flow estimation-based motion compensation for motion segmentation using event camera. The requirement for prior knowledge of the scenario to cluster events into different motion types is addressed by the N-patches optical flow estimation. The framework of motion segmentation combines motion compensation and Pairwise Markov Random Fields, which not only realize the alignment of events in space but also enhance the spatial correlation between events. The proposed approach aims to maximize the contrast of the images of warped events. It does so by minimizing a specially designed energy function that jointly estimates motion parameters and discrete label items. The problem of selecting the optimal value for N∗ when executing N-patches optical flow estimation is discussed in Section 4.2 of the experiments. Finally, the proposed approach is evaluated in several scenes involving high-speed motion, low-light conditions, occlusions, and multiple moving objects. The experimental results show that the proposed approach achieves better segmentation performance than the event-based motion segmentation by motion compensation.

ENLIGHTED phase 3 study: Efficacy and safety of padeliporfin vascular targeted photodynamic therapy (VTP) for treatment of low-grade upper tract urothelial cancer (LG UTUC).

TPS4622 Background: Padeliporfin VTP has demonstrated safety and efficacy for UTUC treatment in a Phase 1 study (NCT03617003). Padeliporfin VTP is a combination product: a drug, padeliporfin administered IV; a device: a laser light delivery system/laser-source of light emitting near-infrared (NIR) light at 753nm; and an optic fiber delivering the light endoluminally to the vicinity of UTUC tumors within the collecting system. We report the preliminary efficacy and safety outcomes of Padeliporfin VTP for Treatment of LG UTUC in ENLIGHTED, a Phase 3 trial (NCT04620239). Methods: This is an open-label phase 3 study from USA, EU and Israel. Key inclusion criteria is up to 2 biopsy-proven LG UTUC with index tumor ≤15mm in the kidney and≤20mm in the ureter with an absence of high-grade cells on cytology. VTP is performed via retrograde upper tract endoscopy under anesthetic and low light conditions, Padeliporfin is injected IV and a laser light diffuser fiber 20-40 mm is positioned in proximity of the tumor through the scope. After Padeliporfin injection, the laser fiber is illuminated for 10 min. The fiber can be repositioned to provide up to 3 treatments within the upper tract. Patients are treated in two phases: Induction(ITP) and Maintenance Treatment Phases (MTP). ITP consists of 1-3 VTPs provided at 4-week intervals until achieving complete response (CR) or treatment failure. Patients achieving CR will proceed to the MTP and be followed with endoscopic evaluation every 3 months with VTP provided for recurrent tumors only up to 12 months. Patients with high-grade recurrences will be deemed treatment failures and removed from the study to standard of care treatment. Patients completing the MTP, will be followed for an additional 48 months for long-term outcomes. Primary outcome is CR on endoscopic evaluation and selective cytology at the time of primary response evaluation (28 ± 3 days post last treatment) during padeliporfin VTP ITP. A total of 100 patients are to be enrolled. As of 21 Jan 2024, 17 patients have been treated. 13 patients completed ITP and had CR 10/13 (77%) and PR 3/13 (23%). Patients baseline characteristics: age 67±12 years; gender: 9(53%) males and 8(47%) females; tumors location: 5(29%) in ureter, 13(76%) in the kidney; number of tumors: 7(41%) 2 tumors, 10(59%) 1 tumor. The most frequent treatment related adverse events were: flank pain 17%, vomiting 8%, fatigue 8%, nausea 6%, hematuria 6%, all Grade 1-2, resolved within few days. Grade 3 serious adverse events 9% (flank pain, hypertension, renal colic, urinary tract infection) were resolved within 2-7 days. To date Padeliporfin VTP has shown evidence of safety and efficacy with preliminary data that is consistent with prior experience. Recruitment for the ENLIGHTED trial is ongoing with results expected to provide basis for approval of a new therapy that clinically benefits pts. Clinical trial information: NCT04620239 .

TQU-HG dataset and comparative study for hand gesture recognition of RGB-based images using deep learning

Hand gesture recognition has great applications in human-computer interaction (HCI), human-robot interaction (HRI), and supporting the deaf and mute. To build a hand gesture recognition model using deep learning (DL) with high results then needs to be trained on many data and in many different conditions and contexts. In this paper, we publish the TQU-HG dataset of large RGB images with low resolution (640×480) pixels, low light conditions, and fast speed (16 fps). TQU-HG dataset includes 60,000 images collected from 20 people (10 male, 10 female) with 15 gestures of both left and right hands. A comparative study with two branches: i) based on Mediapipe TML and ii) Based on convolutional neural networks (CNNs) (you only look once (YOLO); YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLO-Nas, single shot multiBox detector (SSD) VGG16, residual network (ResNet)18, ResNext50, ResNet152, ResNext50, MobileNet V3 small, and MobileNet V3 large), the architecture and operation of CNNs models are also introduced in detail. We especially fine-tune the model and evaluate it on TQU-HG and HaGRID datasets. The quantitative results of the training and testing are presented (F1-score of YOLOv8, YOLO-Nas, MobileNet V3 small, ResNet50 is 98.99%, 98.98%, 99.27%, 99.36%, respectively on the TQU-HG dataset and is 99.21%, 99.37%, 99.36%, 86.4%, 98.3%, respectively on the HaGRID dataset). The computation time of YOLOv8 is 6.19 fps on the CPU and 18.28 fps on the GPU.

The effectiveness of a hybrid MPPT controller based on an artificial neural network and fuzzy logic in low-light conditions

Technological advancement and economic progress have made power consumption a big issue. Concern is growing as traditional energy sources dwindle. In the future, numerous fossil fuels will be insufficient to satisfy human requirements. This motivates research into the feasibility of using renewable energy sources. Renewable energy sources offer a multitude of advantages, including their cost-effectiveness, lack of environmental impact, and sustainable nature. Sunlight is currently the most prevalent source of energy because it is both free and readily accessible. Consequently, photovoltaic (PV) energy is gaining importance in the field of electricity generation. Tracking the maximum power point (MPP) in a solar PV system is challenging due to varying meteorological conditions (irradiance and temperature). To maximise the efficiency of a solar power installation, it is essential to monitor the PV array's optimum power point. This analysis compares the perturb and observe (PO), fuzzy logic (FL), and suggested artificial neural network (ANN)-fuzzy strategy for determining the MPP of a PV system with minimal radiation exposure. Simulation results show that at low irradiation levels, the proposed ANN-fuzzy maximum power point tracking (MPPT) unit controller is superior to the FL and PO MPPT controllers in terms of tracking maximum power.

Splitting light pollution: Wavelength effects on the activity of two sandy beach species

Artificial Light at Night (ALAN) threatens to disrupt most natural habitats and species, including those in coastal settings, where a growing number of studies have identified ALAN impacts. A careful examination of the light properties behind those impacts is important to better understand and manage the effects of this stressor. This study focused on ALAN monochromatic wavelengths and examined which types of light spectra altered the natural activity of two prominent coastal species from the Pacific southeast: the talitroid amphipod Orchestoidea tuberculata and the oniscoid isopod Tylos spinulosus. We compared the natural daylight/night activity of these organisms with the one they exhibit when exposed to five different ALAN wavelengths: lights in the violet, blue, green, amber, and red spectra. Our working hypothesis was that ALAN alters these species’ activity at night, but the magnitude of such impact differs depending on light wavelengths. Measurements of activity over 24 h cycles for five consecutive days and in three separate experiments confirmed a natural circadian activity pattern in both species, with strong activity at night (∼90% of probability) and barely any activity during daylight. However, when exposed to ALAN, activity declined significantly in both species under all light wavelengths. Interestingly, amphipods exhibited moderate activity (∼40% of probability) when exposed to red lights at night, whereas isopods shifted some of their activity to daylight hours in two of the experiments when exposed to blue or amber lights, suggesting a possible alteration in this species circadian rhythm. Altogether, our results were consistent with our working hypothesis, and suggest that ALAN reduces night activity, and some wavelengths have differential effects on each species. Differences between amphipods and isopods are likely related to their distinct adaptations to natural low-light habitat conditions, and therefore distinct sensitivity to ALAN.

Does water transparency control the banding in shallow water iron formations?

Abstract Iron formations (IFs) are marine chemical sediments that are conined to the Precambrian rock record and provide unique insights into the co-evolution of the atmosphere-hydrosphere and biosphere through almost three billion years of Earth’s history. IFs commonly appear throughout the Archaean until the Palaeoproterozoic ca. 1.8 billion years ago and re-appear during the “Snowball Earth” epoch in the Neoproterozoic. The formation and deposition mechanism(s) of IFs are, however, still incompletely understood, hindering unique interpretations of palaeoenvironments. Many IFs are banded iron formations (BIFs) with layer thickness of alternating Fe- and Si-rich layers ranging over several orders of magnitude from the nanometre to the metre scale. A second textural type, so called granular iron formations (GIFs) that form above storm wave base become widespread in the Palaeoproterozoic. Understanding the fundamental mechanisms that are responsible for the textural types and the periodicity of banding in BIFs is crucial to link these features to the environmental and geochemical evolution of the Earth. We here provide a conceptual model that highlights the role of changing light conditions and water transparency for Iron Formation (IF) precipitation. We show that the model is particularly feasible for IFs deposited in shallow waters but may also be applicable for some IFs deposited in deeper water settings. The model builds on other primary Banded Iron Formation (BIF) precipitation models postulating that Fe(III)-(oxyhydr)oxide production can be dominated by anoxygenic photoautotrophic Fe2+-oxidising bacteria. These so called photoferrotrophs are adapted to very low light levels corresponding to about 1% of the light level required by oxygen-producing phototrophs allowing them to thrive deep down in the water column. The depth of Fe(III)-(oxyhydr)oxide production is mainly controlled by water turbidity which controls how deep photosynthetically available radiation (PAR) penetrates the water column. Eutrophic conditions result in relatively shallow Fe(III)-(oxyhydr)oxide production depth due to turbidity either induced by the biomass itself and/or by particles that are actively or passively produced by microorganisms (e.g., Fe(III) and Mn(IV)-(oxyhydr)oxides, sulphides), triggering the formation of Fe-rich bands. During oligotrophic stages, Fe(III)-(oxyhydr)oxides are only produced relatively deep down in the water column, so that only silica-rich bands form in the Fe(III)-(oxyhydr)oxide free upper water column. Reactive transport modelling adopted from Ozaki et al., (2019) shows that besides upwelling and nutrient supply, alternating Fe(III) production depth is mainly associated with changing light conditions as a result of water transparency. Periodicities reflected by alternating Fe- and Si-rich bands in IFs in our model can thus be associated with: (1) nutrient supply patterns; (2) additional sources of turbidity in the water column such as Fe(III) and Mn(IV) oxide particles, sulphides, and wind-blown silicate particles; or (3) formation and clearing of organic haze in the atmosphere. One or all of these reasons for low light conditions seem to become more important in the Palaeoproterozoic (<2.4 Ga) and could be partly responsible for the more widespread occurrences of shallow marine granular IFs relative to former epochs, which is often assigned to the gradual oxidation of the ocean. Our model shows that variable water transparency should be considered as additional factor for IF deposition especially for shallow marine settings. This model also reasonably explains the prominent layering in BIFs as syn-depositional feature.

Pixel-wise low-light image enhancement based on metropolis theorem

Images taken in low-light conditions frequently encounter visibility problems, such as severe noise, reduced brightness, and low contrast. This paper introduces an approach to enhance low-light images using the Metropolis Theorem (MT). The method begins by applying a global gamma correction to the input image, followed by transforming the globally corrected image into the HSV (Hue, Saturation, Value - V) domain. To achieve multi-scale decomposition, an application of the MT is proposed, resulting in approximation and detail sub-images of the V component. Subsequently, local gamma correction is employed on both the final approximation and detail images to enhance local contrast. The refined approximation and detail images are then combined to reconstruct the refined V component. The reconstructed image is obtained by weighting each band of the image with the refined V component. Experimental results demonstrate that the proposed method outperforms state-of-the-art methods, providing improved visual quality and more natural colors.

A multi-branch dual attention segmentation network for epiphyte drone images

Acquiring images of epiphytes growing on trees using Unmanned Aerial Vehicles (UAVs) enables botanists to efficiently collect data on these important plant species. Despite the advantages offered by UAVs, challenges such as complex backgrounds, uneven lighting inside the tree canopy, and accessibility issues hinder the acquisition of quality images, resulting in acquiring images datasets of heterogenous quality. AI/Deep Learning algorithms can be used to segment target plants in these images for selecting sampling locations. Existing DL models require large volume of data for training, and they tend to prioritize local features over global ones, impacting segmentation accuracy, particularly on smaller, heterogeneous quality image datasets. To overcome these limitations, we propose a multi-branch dual attention segmentation network designed to effectively handle small datasets with heterogeneous quality. The proposed network incorporates dedicated branches for extracting both global and local features, utilizing spatial and channel attention mechanisms to focus on important regions. Through a fusion process and a decoder with crossed fusion technique, this network effectively combines and enhances features from multiple branches, resulting in improved segmentation performance. Output obtained from the trained model demonstrated major improvements in predicting the boundary regions and class labels, even in close-range, low-light, and zoomed/cropped images. The average Intersection over Union (IoU) scores of the trained model was 5% higher for images acquired close range, 48% higher for images in low-light conditions, and 68% higher for zoomed/cropped images when compared to those obtained from TransUnet, a state-of-the-art vision transformer model trained on epiphyte dataset. The proposed network can be used for segmenting epiphytes in images of heterogeneous quality as well as identifying targets in images acquired in domains such as agriculture and forestry.

The recent disappearance of a persistent Planktothrix bloom: Characterization of a regime shift in the phytoplankton of Sandusky Bay (USA)

Sandusky Bay is the drowned mouth of the Sandusky River in the southwestern portion of Lake Erie. The bay is a popular recreation location and a regional source for drinking water. Like the western basin of Lake Erie, Sandusky Bay is known for being host to summer cyanobacterial harmful algal blooms (cHABs) year after year, fueled by runoff from the predominantly agricultural watershed and internal loading of legacy nutrients (primarily phosphorus). Since at least 2003, Sandusky Bay has harbored a microcystin-producing bloom of Planktothrix agardhii, a species of filamentous cyanobacteria that thrives in low light conditions. Long-term sampling (2003–2018) of Sandusky Bay revealed regular Planktothrix-dominated blooms during the summer months, but in recent years (2019–2022), 16S rRNA gene community profiling revealed that Planktothrix has largely disappeared. From 2017–2022, microcystin decreased well below the World Health Organization (WHO) guidelines. Spring TN:TP ratios increased in years following dam removal, yet there were no statistically significant shifts in other physicochemical variables, such as water temperature and water clarity. With the exception of the high bloom of Planktothrix in 2018, there was no statistical difference in chlorophyll during all other years. Concurrent with the disappearance of Planktothrix, Cyanobium spp. have become the dominant cyanobacterial group. The appearance of other potential toxigenic genera (i.e., Aphanizomenon, Dolichospermum, Cylindrospermopsis) may motivate monitoring of new toxins of concern in Sandusky Bay. Here, we document the regime shift in the cyanobacterial community and propose evidence supporting the hypothesis that the decline in the Planktothrix bloom was linked to the removal of an upstream dam on the Sandusky River.

Functional Algorithm Model Mumford-Shah for Facial Recognition and Image Segmentation in Low Light Conditions, Lima 2023

Functional Algorithm Model Mumford-Shah for Facial Recognition and Image Segmentation in Low Light Conditions, Lima 2023

Drowsiness Detection and Alert System

Drowsiness detection is a solution for identifying signs of fatigue or sleepiness in individuals. One of the key features of our model is that it can detect drowsiness at night as well using Mobile cameras (infrared sensors). The system captures infrared images of the person's face and analyzes the physiological and behavioral cues related to drowsiness. Infrared sensors allow for drowsiness detection in low-light conditions, making it particularly useful for night-time scenarios such as night driving. The system can trigger alerts or interventions if drowsiness is detected, helping to prevent accidents or mistakes. We will be using libraries like OpenCV, TensorFlow, CNN, and VGG19 features in our model. By combining the accessibility of Android devices with the advanced capabilities of the Deep Learning algorithm, drowsiness detection using infrared sensors has the potential to greatly improve the safety and productivity of individuals in their daily lives.