Similar Color Research Articles

Navigation path extraction for inter-row robots in Panax notoginseng shade house based on Im-YOLOv5s.

The accurate extraction of navigation paths is crucial for the automated navigation of agricultural robots. Navigation line extraction in complex environments such as Panax notoginseng shade house can be challenging due to factors including similar colors between the fork rows and soil, and the shadows cast by shade nets. In this paper, we propose a new method for navigation line extraction based on deep learning and least squares (DL-LS) algorithms. We improve the YOLOv5s algorithm by introducing MobileNetv3 and ECANet. The trained model detects the seven-fork roots in the effective area between rows and uses the root point substitution method to determine the coordinates of the localization base points of the seven-fork root points. The seven-fork column lines on both sides of the plant monopoly are fitted using the least squares method. The experimental results indicate that Im-YOLOv5s achieves higher detection performance than other detection models. Through these improvements, Im-YOLOv5s achieves a mAP (mean Average Precision) of 94.9%. Compared to YOLOv5s, Im-YOLOv5s improves the average accuracy and frame rate by 1.9% and 27.7%, respectively, and the weight size is reduced by 47.9%. The results also reveal the ability of DL-LS to accurately extract seven-fork row lines, with a maximum deviation of the navigation baseline row direction of 1.64°, meeting the requirements of robot navigation line extraction. The results shows that compared to existing models, this model is more effective in detecting the seven-fork roots in images, and the computational complexity of the model is smaller. Our proposed method provides a basis for the intelligent mechanization of Panax notoginseng planting.

Introgression Underlies Phylogenetic Uncertainty But Not Parallel Plumage Evolution in a Recent Songbird Radiation.

Instances of parallel phenotypic evolution offer great opportunities to understand the evolutionary processes underlying phenotypic changes. However, confirming parallel phenotypic evolution and studying its causes requires a robust phylogenetic framework. One such example is the "black-and-white wagtails," a group of 5 species in the songbird genus Motacilla: 1 species, Motacilla alba, shows wide intra-specific plumage variation, while the 4r others form 2 pairs of very similar-looking species (M. aguimp + M. samveasnae and M. grandis + M. maderaspatensis, respectively). However, the 2 species in each of these pairs were not recovered as sisters in previous phylogenetic inferences. Their relationships varied depending on the markers used, suggesting that gene tree heterogeneity might have hampered accurate phylogenetic inference. Here, we use whole genome resequencing data to explore the phylogenetic relationships within this group, with a special emphasis on characterizing the extent of gene tree heterogeneity and its underlying causes. We first used multispecies coalescent methods to generate a "complete evidence" phylogenetic hypothesis based on genome-wide variants, while accounting for incomplete lineage sorting (ILS) and introgression. We then investigated the variation in phylogenetic signal across the genome to quantify the extent of discordance across genomic regions and test its underlying causes. We found that wagtail genomes are mosaics of regions supporting variable genealogies, because of ILS and inter-specific introgression. The most common topology across the genome, supporting M. alba and M. aguimp as sister species, appears to be influenced by ancient introgression. Additionally, we inferred another ancient introgression event, between M. alba and M. grandis. By combining results from multiple analyses, we propose a phylogenetic network for the black-and-white wagtails that confirms that similar phenotypes evolved in non-sister lineages, supporting parallel plumage evolution. Furthermore, the inferred reticulations do not connect species with similar plumage coloration, suggesting that introgression does not underlie parallel plumage evolution in this group. Our results demonstrate the importance of investing genome-wide patterns of gene tree heterogeneity to help understand the mechanisms underlying phenotypic evolution. [Gene tree heterogeneity; incomplete lineage sorting; introgression; parallel evolution; phylogenomics; plumage evolution; wagtails.].

Rates and properties of Type Ia supernovae in galaxy clusters within the dark energy survey

ABSTRACT We identify 66 photometrically classified Type Ia supernovae (SNe Ia) from the Dark Energy Survey (DES) that have occurred within red-sequence selected galaxy clusters. We compare light-curve and host galaxy properties of the cluster SNe to 1024 DES SNe Ia located in field galaxies, the largest comparison of two such samples at high redshift (z > 0.1). We find that cluster SN light curves decline faster than those in the field (97.7 per cent confidence). However, when limiting these samples to host galaxies of similar colour and mass, there is no significant difference in the SN light-curve properties. Motivated by previous detections of a higher-normalized SN Ia delay-time distribution in galaxy clusters, we measure the intrinsic rate of SNe Ia in cluster and field environments. We find the average ratio of the SN Ia rate per galaxy between high-mass ($10\le \log \mathrm{(\mathit{ M}_{*}/{\rm \mathit{ M}}_{\odot })} \le 11.25$) cluster and field galaxies to be 0.594 ± 0.068. This difference is mass-dependent, with the ratio declining with increasing mass, which suggests that the stellar populations in cluster hosts are older than those in field hosts. We show that the mass-normalized rate (or SNe per unit mass) in massive–passive galaxies is consistent between cluster and field environments. Additionally, both of these rates are consistent with rates previously measured in clusters at similar redshifts. We conclude that in massive–passive galaxies, which are the dominant hosts of cluster SNe, the cluster delay-time distribution is comparable to the field.

Cost-effective pearlescent pigments with high near-infrared reflectance and outstanding energy-saving ability for mitigating urban heat island effect

Cooling down the surface temperature of buildings and saving energy consumption are tasks of top priority in mitigating the urban heat island (UHI) effect. The bright and colorful pearlescent pigmented cool coatings with high near-infrared (NIR) reflectance possess promise in enhancing the reflective effect of solar light and the visual aspect, which is still an undeveloped domain. Herein, these novel solar reflective coatings containing mediums and pearlescent pigments with high NIR reflectance of 85% embodied obvious energy-saving effects on building materials. EnergyPlus software was used to simulate the energy-saving effect of four applied climate zones, including New York, Nanchang, Lima, and Brasilia. Compared to commercial pigments (Fe2O3, γ-Ce2S3) with similar colors, the RK 402 and RK 301 pearlescent pigments demonstrate superior energy-saving capabilities in Nanchang climate zone, and can save 1.00 kW h m−2 and 1.60 kW h m−2 annually, respectively. In the warmer climate zone of Brasilia, the highest energy-saving efficiency is 2.06 kW h m−2 over one year by RK 301 pearlescent pigment. After irradiated by an NIR lamp, the surface temperature of the RK 301 coating can be reduced by 34.8 °C compared with that of the Fe2O3 commercial coating, and that of RK 402 coating is 5.2 °C compared with the γ-Ce2S3 commercial coating. Furthermore, the 7-day field test also has consistent effect of cooling down the surface temperature. This work demonstrated that pearlescent coatings with brilliant colors have great potential for application as cold materials in building areas to mitigate the UHI effect.

Predicting the Yields of z > 6.5 Quasar Surveys in the Era of Roman and Rubin

About 70 luminous quasars discovered at z > 6.5 are strongly biased toward the bright end, thus not providing a comprehensive view of quasar abundance beyond the cosmic dawn. We present the predicted results of the Roman/Rubin high-redshift quasar survey, yielding 3 times more, 2–4 mag deeper quasar samples, probing high-redshift quasars across a broad range of luminosities, especially faint quasars at L bol ∼ 1010 L ⊙ or M 1450 ∼ −22, which are currently poorly explored. We include high-z quasars, galactic dwarfs, and low-z compact galaxies with similar colors as quasar candidates. We create mock catalogs based on population models to evaluate selection completeness and efficiency. We utilize the classical color dropout method in the z and Y bands to select primary quasar candidates, followed up with the Bayesian selection method to identify quasars. We show that overall selection completeness >80% and efficiency ∼10% at 6.5 < z < 9, with 180 quasars at z > 6.5, 20 at z > 7.5, and 2 at z > 8.5. The quasar yields depend sensitively on the assumed quasar luminosity shape and redshift evolution. Brown dwarf rejection through proper motion up to 50% can be made for stars brighter than 25 mag, low-z galaxies dominate at fainter magnitude. Our results show that Roman/Rubin are able to discover a statistical sample of the earliest and faintest quasars in the Universe. The new valuable data sets are worth follow-up studies with JWST and Extremely Large Telescopes to determine the quasar luminosity function faint end slope and constraint the supermassive black holes growth in the early Universe.

조지아오키프 작품을 모티브로 한 환타지메이크업 일러스트레이션에 관한 연구

The purpose of this paper is to pursue a basic foundation that can expand the creative expression area of fantasy makeup, which resumed work by repaying abstract designs using the shipping rights and colors of motifs in Georgia O'Keeffe's work. do. The research method consisted of work research and empirical research. In the backpack study, the rights of fantasy makeup and the life and life of GyauKeefe were examined through origin research and professional data. As a result of the study, the black wavy lines of the roots of the trees seen in Gle O'Keefe's Autumn Tree-The Maple work were connected to emphasize, transform, or express the sense of life. In the Red Canna piece, it meets the repeating colors of the petals and starts getting brighter. In the Music-Pink and Blue work, similar colors and wavy lines were harmonized as if negotiating in order to emphasize the expanding activation by demonstrating musical rhythm.

3D유닛패턴 체계도 기반의 장식용 아트타일 개발과 호환성 검토

This study developed basic units such as cells and units based on the FDP design methodology with the product scalability of flexible design patterns (FDP) that simultaneously evoke images of order and change in mind, and verified the compatibility of design by combining cells of different shapes. The cells developed for verification are largely divided into straight lines and curves, lines and area shapes, and the colors are black, white, and yellow tricolors, which are combined with the same color, similar color, and contrast color. Combination methods were divided into isomorphism, similarity, and heteromorphic, and examined in connection with color. The deployment method was performed by comparing the regular deployment (configuration) of cells and units with the irregular deployment (configuration). As a result, most of the combinations of cells with different shapes were found to have visual effects of order, change, and unity. Above all, it was confirmed that the combination of shapes has a strong characteristic of inducing unity and harmony, and the contrast between black and white has a side that triggers a change.On the other hand, the visual effect of FDP is possible because of the division method based on golden division, the enhancement of connectivity using diagonal and connection pointer, and the symmetric manipulation. This formative technique is expected to be used as educational contents for experiential learning of basic molding in connection with 3D printing technology.

The environmental footprint of Holocene societies: a multi-temporal study of trails in the Judean Desert, Israel

The global distribution of footpaths and their inferred antiquity implies that they are widespread spatial and temporal anthropogenic landscape units. Arid environments are of special interest for investigating historically used footpaths, as older routes may preserve better due to minimal modern impact and slower pedogenic processes. Here we examine footpaths in the Judean Desert of the southern Levant, a human hotspot throughout the Holocene. We studied one modern and two archaeological footpaths (one attributed to the Early Bronze Age and one to the Roman period) using micromorphology, bulk samples laboratory analysis, and remote sensing. Field observations and color analysis indicate that footpaths in the studied arid limestone environment can result in brighter surface color than their non-path surroundings. Similar color changes are reflected using both laboratory analysis and high-resolution remote sensing, where the difference is also significant. Microscopically, the footpaths studied tend to be less porous and with fewer biogenic activities when compared to their non-path controls. However, the two ancient footpaths studied do exhibit minimal indicators of biogenic activities that are not detectable in the modern footpath sample. Our study shows that high-resolution remote sensing coupled with micromorphology, while using appropriate local modern analogies, can help to locate and assess both the environmental effect and the antiquity of footpaths.

A tea bud segmentation, detection and picking point localization based on the MDY7-3PTB model.

The identification and localization of tea picking points is a prerequisite for achieving automatic picking of famous tea. However, due to the similarity in color between tea buds and young leaves and old leaves, it is difficult for the human eye to accurately identify them. To address the problem of segmentation, detection, and localization of tea picking points in the complex environment of mechanical picking of famous tea, this paper proposes a new model called the MDY7-3PTB model, which combines the high-precision segmentation capability of DeepLabv3+ and the rapid detection capability of YOLOv7. This model achieves the process of segmentation first, followed by detection and finally localization of tea buds, resulting in accurate identification of the tea bud picking point. This model replaced the DeepLabv3+ feature extraction network with the more lightweight MobileNetV2 network to improve the model computation speed. In addition, multiple attention mechanisms (CBAM) were fused into the feature extraction and ASPP modules to further optimize model performance. Moreover, to address the problem of class imbalance in the dataset, the Focal Loss function was used to correct data imbalance and improve segmentation, detection, and positioning accuracy. The MDY7-3PTB model achieved a mean intersection over union (mIoU) of 86.61%, a mean pixel accuracy (mPA) of 93.01%, and a mean recall (mRecall) of 91.78% on the tea bud segmentation dataset, which performed better than usual segmentation models such as PSPNet, Unet, and DeeplabV3+. In terms of tea bud picking point recognition and positioning, the model achieved a mean average precision (mAP) of 93.52%, a weighted average of precision and recall (F1 score) of 93.17%, a precision of 97.27%, and a recall of 89.41%. This model showed significant improvements in all aspects compared to existing mainstream YOLO series detection models, with strong versatility and robustness. This method eliminates the influence of the background and directly detects the tea bud picking points with almost no missed detections, providing accurate two-dimensional coordinates for the tea bud picking points, with a positioning precision of 96.41%. This provides a strong theoretical basis for future tea bud picking.

Survey of gravitationally lensed objects in HSC imaging (SuGOHI) – IX. Discovery of strongly lensed quasar candidates

ABSTRACT We report the discovery of new lensed quasar candidates in the imaging data of the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP) Data Release 4, covering $1\, 310\deg ^2$ of the sky with median seeing in the i band of ≈0.6 arcsec. In addition to two catalogues of Million Quasars Catalog v7.2 and AllWISE Catalog of Mid-Infra Red Active Galactic Nuclei, containing confirmed and candidate quasars, we preselect quasar sources using colour cuts from the HSC (grizy) and unWISE (W1 + W2) photometric data based on SDSS spectroscopic catalogues. We search for the presence of multiple point sources with similar colour through the convolution of the Laplacian of the preselected quasar image cutouts with the Laplacian of the point spread function, resulting in a reduction of lens candidates from 1 652 329 to 121 511 (7.4 per cent). After visual binary classification, we grade 6 199 (0.4 per cent) potential lenses on a scale of 0 to 3, with 3 indicating a lens and 0 indicating a non-lens. Finally we obtain 162 lens candidates with an average grade of ≥2, and among them, we successfully recover 18 known lenses. By fitting the light distribution and removing the known contaminants, we discover that 57 new systems contain at least two point sources and a galaxy in between, including 10 possible quadruply lensed quasars. This new sample exhibits a median separation of 1.26 arcsec and a magnitude limit of i ≈ 22. Spectroscopic or high-resolution imaging follow up on these newly discovered lensed quasar candidates will further allow their natures to be confirmed.

Fuzzy C‐means clustering algorithm based on superpixel merging and multi‐feature adaptive fusion measurement

AbstractThe fuzzy C‐means clustering (FCM) algorithm is widely used in greyscale and colour image segmentation, especially in real colour images. However, in the process of interested regions extraction, it performs barely satisfactory due to the use of single distance metric in traditional FCM. In order to address the issue, a fuzzy C‐means clustering algorithm based on superpixel merging and multi‐feature adaptive fusion measurement (FCM‐SM) are proposed. First of all, since superpixel image adapts well to irregular image boundaries, the FCM algorithm is elevated to the superpixel level. Moreover, the superpixel merging module based on colour similarity is introduced to improve the traditional watershed transformation by incorporating more gradient and intensity information. Finally, the multi‐feature adaptive fusion measurement is designed in the clustering process to incorporate multiple sources of spatial information, which comprehensively measures the information from different fields to enhance the segmentation capability further. Experiments performed on Berkeley Segmentation Dataset and Brain Tumor Segmentation Dataset demonstrate that the proposed algorithm provides better segmentation results than benchmarks during the object extraction.

Corn-Based Gluten-Free Snacks Supplemented with Various Dried Fruits: Characteristics of Physical Properties and Effect of Variables

The market of extruded products is constantly growing and the incorporation of fruit items into their recipe, can made a crisp snack product a healthy one of acceptable flavor. The subject of this work is the evaluation of the effects of production screw speed, fruit type and amount on selected physical properties (expansion index, bulk density, water absorption and solubility, texture profile and color balance) of corn-based gluten-free crisps supplemented with various amounts (0–20%) of dried fruits (apple, white mulberry, goji berry, elderberry, blackberry) processed at variable screw speeds (80, 100 and 120 rpm). This work demonstrates that it is possible to obtain marketable extruded snacks with natural color coming from the incorporated dried fruits and with adequate expansion and texture if addition was up to 10% of all the tested fruits. Moreover, very good aeration, crispy texture and acceptable natural color was found if dried elderberry and blackberry were added to snacks even at 15 and 20%. Application of 15 or 20% of apple, white mulberry and goji berries showed similar color profiles and caused decrease in texture and expansion of snacks. The rotational screw speed effect differs significantly only in hardness and cutting force of the supplemented corn crisps.

Underwater Fish Segmentation Algorithm Based on Improved PSPNet Network.

With the sustainable development of intelligent fisheries, accurate underwater fish segmentation is a key step toward intelligently obtaining fish morphology data. However, the blurred, distorted and low-contrast features of fish images in underwater scenes affect the improvement in fish segmentation accuracy. To solve these problems, this paper proposes a method of underwater fish segmentation based on an improved PSPNet network (IST-PSPNet). First, in the feature extraction stage, to fully perceive features and context information of different scales, we propose an iterative attention feature fusion mechanism, which realizes the depth mining of fish features of different scales and the full perception of context information. Then, a SoftPool pooling method based on fast index weighted activation is used to reduce the numbers of parameters and computations while retaining more feature information, which improves segmentation accuracy and efficiency. Finally, a triad attention mechanism module, triplet attention (TA), is added to the different scale features in the golden tower pool module so that the space attention can focus more on the specific position of the fish body features in the channel through cross-dimensional interaction to suppress the fuzzy distortion caused by background interference in underwater scenes. Additionally, the parameter-sharing strategy is used in this process to make different scale features share the same learning weight parameters and further reduce the numbers of parameters and calculations. The experimental results show that the method presented in this paper yielded better results for the DeepFish underwater fish image dataset than other methods, with 91.56% for the Miou, 46.68 M for Params and 40.27 G for GFLOPS. In the underwater fish segmentation task, the method improved the segmentation accuracy of fish with similar colors and water quality backgrounds, improved fuzziness and small size and made the edge location of fish clearer.

Remote Sensing Image Haze Removal Based on Superpixel

The presence of haze significantly degrades the quality of remote sensing images, resulting in issues such as color distortion, reduced contrast, loss of texture, and blurred image edges, which can ultimately lead to the failure of remote sensing application systems. In this paper, we propose a superpixel-based visible remote sensing image dehazing algorithm, namely SRD. To begin, the remote sensing haze images are divided into content-aware patches using superpixels, which cluster adjacent pixels considering their similarities in color and brightness. We assume that each superpixel region shares the same atmospheric light and transmission properties. Subsequently, methods to estimate local atmospheric light and transmission within each superpixel are proposed. Unlike existing dehazing algorithms that assume a globally constant atmospheric light, our approach considers the global heterogeneous distribution of the atmospheric ambient light, which allows us to model it as a global non-uniform variable. Furthermore, we introduce an effective atmospheric light estimation method inspired by the maximum reflectance prior. Moreover, recognizing the wavelength-dependent nature of light transmission, we independently estimate the transmittance for each RGB channel of the input image. The quantitative and qualitative evaluation results of comprehensive experiments on synthetic datasets and real-world samples demonstrate the superior performance of the proposed algorithm compared to state-of-the-art methods for remote sensing image dehazing.

Sustainable dyeing of mulberry silk fabric using extracts of green tea (Camellia sinensis): Extraction, mordanting, dyed silk fabric properties and silk-dye interaction mechanism

Sustainability and environmental awareness had an impulse in increasing the global interest to utilize the industrial crops for producing new colouring shades in place of hazardous synthetic dyes. The present research focuses on extracting functional components from green tea (Camelia sinensis) leaves using water and methanol/water solvent systems and their application in the mordant dyeing of mulberry silk fabrics. The effect of pH was studied to maximize the yield of total polyphenols, flavonoids, and antioxidant properties of the solvent extracts. Compared to the unmordanted dyed silk, the mordant dyeing of silk with tea extracts at a temperature of 80 ℃ for 45 min resulted in significant improvements in various color properties. These enhancements include the shade, color strength (K/S), color coordinates, and fastness properties of the dyed silk fabric. Mordant dyeing of silk fabric enhanced UV protection and antimicrobial activities as well. The bonding properties of tea polyphenols on silk was well demonstrated with the help of X-ray photoelectron spectroscopy (XPS). To establish the mechanism of dyeing, catechin was used as a standard and mordant dyeing in silk, resulting in similar golden-yellow coloration. It shows identical properties with the silk dyed with tea extracts resulting in the fact that catechins are the responsible components in silk dyeing while using tea extracts. Theoretical interaction energy (Einteraction = −13.836 kJ/mol) calculation established good stability of mordanted catechin dyeing on silk, and catechin stability in different solvents were in the order catechin-water > catechin-methanol > catechin-ethanol. This work suggests that the mulberry silk fabrics dyed with green tea extract showed diverse color variations, which is highly promising to be used as a natural dye in the textile industry.

A new spurless species of Impatiens sect. Microsepalae (Balsaminaceae) from south India

An interesting balsam with spurless lower sepal is described here as a new species to the Impatiens Sect. Microsepalae (Balsaminaceae) from southern parts of Western Ghats, India. Detailed notes on description, ecology, and distribution are provided along with a comparison of its morphologically similar species and colour photographs.

Intelligent Detection of Lightweight “Yuluxiang” Pear in Non-Structural Environment Based on YOLO-GEW

To detect quickly and accurately “Yuluxiang” pear fruits in non-structural environments, a lightweight YOLO-GEW detection model is proposed to address issues such as similar fruit color to leaves, fruit bagging, and complex environments. This model improves upon YOLOv8s by using GhostNet as its backbone for extracting features of the “Yuluxiang” pears. Additionally, an EMA attention mechanism was added before fusing each feature in the neck section to make the model focus more on the target information of “Yuluxiang” pear fruits, thereby improving target recognition ability and localization accuracy. Furthermore, the CIoU Loss was replaced with the WIoUv3 Loss as the loss function, which enhances the capability of bounding box fitting and improves model performance without increasing its size. Experimental results demonstrated that the enhanced YOLO-GEW achieves an F1 score of 84.47% and an AP of 88.83%, while only occupying 65.50% of the size of YOLOv8s. Compared to lightweight algorithms such as YOLOv8s, YOLOv7-Tiny, YOLOv6s, YOLOv5s, YOLOv4-Tiny, and YOLOv3-Tiny; there are improvements in AP by 2.32%, 1.51%, 2.95%, 2.06%, 2.92%, and 5.38% respectively. This improved model can efficiently detect “Yuluxiang” pears in non-structural environments in real-time and provides a theoretical basis for recognition systems used by picking robots.

Structure-preserving feature alignment for old photo colorization

Deep learning techniques have made significant advancements in reference-based colorization by training on large-scale datasets. However, directly applying these methods to the task of colorizing old photos is challenging due to the lack of ground truth and the notorious domain gap between natural gray images and old photos. To address this issue, we propose a novel CNN-based algorithm called SFAC, i.e., Structure-preserving Feature Alignment Colorizer. SFAC is trained on only two images for old photo colorization, eliminating the reliance on big data and allowing direct processing of the old photo itself to overcome the domain gap problem. Our primary objective is to establish semantic correspondence between the two images, ensuring that semantically related objects have similar colors. We achieve this through a feature distribution alignment loss that remains robust to different metric choices. However, utilizing robust semantic correspondence to transfer color from the reference to the old photo can result in inevitable structure distortions. To mitigate this, we introduce a structure-preserving mechanism that incorporates a perceptual constraint at the feature level and a frozen-updated pyramid at the pixel level. Extensive experiments demonstrate the effectiveness of our method for old photo colorization, as confirmed by qualitative and quantitative metrics.

Relationships of the gall midge, Asphondylia miki Wachtl (Diptera, Cecidomyiidae), with the host plant, alfalfa (Medicago sativa) and ectoparasites of its larvae and pupae (Hymenoptera, Eulophidae) in the forest-steppe of the middle volga region

The studies were carried out in crops of alfalfa ( Medicago sativa ssp. × varia ) in 2018-2022 in the forest-steppe of Samara Province. Asphondylia miki galls were taken into account in alfalfa crops from the beginning of their formation to the end of the growing season on sites from 0.1 to 1 m2 in 10-fold repetition. Statistical processing of the obtained data was carried out by methods of variance analysis. At spring sowing of alfalfa, as well as in its perennial crops, gall midge A. miki develops in two, and during summer sowing in the rst year of the culture’s life in one generation per year, pupae overwinter in galls, the imago of the wintering generation occurs in April. The greatest damage to alfalfa by the larvae of A. miki (40-68% of generative shoots) and the number of its galls (500-830 copies/m2) were observed in the rst year of the culture’s life when it was sown in summer, with the coincidence of the greatest ight of imago with the budding of the culture, which led to signi cant losses in the yield of alfalfa seeds (13-81%). In summer alfalfa crops of the rst year of development without the use of chemical plant protection agents, parasitic hymenopterans contribute to the death of larvae and pupae of A. miki in on average 60% of its galls. Among them, a gregarious ectoparasite Sigmophora brevicornis (Eulophidae) mainly preying on A. miki larvae accounted for about 77% of the individuals of the identi ed parasites. In the S. brevicornis population, males and females had similar body coloration and belonged to three phenotypes: the lightest brownish yellow, yellowish brown, and dark coloration, which accounted for 53, 28 and 19% of parasite individuals, respectively.

Divergences in color perception between deep neural networks and humans

Deep neural networks (DNNs) are increasingly proposed as models of human vision, bolstered by their impressive performance on image classification and object recognition tasks. Yet, the extent to which DNNs capture fundamental aspects of human vision such as color perception remains unclear. Here, we develop novel experiments for evaluating the perceptual coherence of color embeddings in DNNs, and we assess how well these algorithms predict human color similarity judgments collected via an online survey. We find that state-of-the-art DNN architectures — including convolutional neural networks and vision transformers — provide color similarity judgments that strikingly diverge from human color judgments of (i) images with controlled color properties, (ii) images generated from online searches, and (iii) real-world images from the canonical CIFAR-10 dataset. We compare DNN performance against an interpretable and cognitively plausible model of color perception based on wavelet decomposition, inspired by foundational theories in computational neuroscience. While one deep learning model — a convolutional DNN trained on a style transfer task — captures some aspects of human color perception, our wavelet algorithm provides more coherent color embeddings that better predict human color judgments compared to all DNNs we examine. These results hold when altering the high-level visual task used to train similar DNN architectures (e.g., image classification versus image segmentation), as well as when examining the color embeddings of different layers in a given DNN architecture. These findings break new ground in the effort to analyze the perceptual representations of machine learning algorithms and to improve their ability to serve as cognitively plausible models of human vision. Implications for machine learning, human perception, and embodied cognition are discussed.