Map Scale Research Articles

Risky business: human-related data is lacking from Lyme disease risk models.

Used as a communicative tool for risk management, risk maps provide a service to the public, conveying information that can raise risk awareness and encourage mitigation. Several studies have utilized risk maps to determine risks associated with the distribution of Borrelia burgdorferi, the causal agent of Lyme disease in North America and Europe, as this zoonotic disease can lead to severe symptoms. This literature review focused on the use of risk maps to model distributions of B. burgdorferi and its vector, the blacklegged tick (Ixodes scapularis), in North America to compare variables used to predict these spatial models. Data were compiled from the existing literature to determine which ecological, environmental, and anthropic (i.e., human focused) variables past research has considered influential to the risk level for Lyme disease. The frequency of these variables was examined and analyzed via a non-metric multidimensional scaling analysis to compare different map elements that may categorize the risk models performed. Environmental variables were found to be the most frequently used in risk spatial models, particularly temperature. It was found that there was a significantly dissimilar distribution of variables used within map elements across studies: Map Type, Map Distributions, and Map Scale. Within these map elements, few anthropic variables were considered, particularly in studies that modeled future risk, despite the objective of these models directly or indirectly focusing on public health intervention. Without including human-related factors considering these variables within risk map models, it is difficult to determine how reliable these risk maps truly are. Future researchers may be persuaded to improve disease risk models by taking this into consideration.

LF-YOLOv4: a lightweight detection model for enhancing the fusion of image features of surface defects in lithium batteries

In response to the low accuracy of traditional methods for detecting surface defects in lithium batteries, as well as the problems of large model size and high computational complexity in current detection models, this article proposes a new lightweight LF-YOLOv4 model that enhances image feature fusion. Firstly, replace the CSPDarknet53 backbone network in YOLOv4 with a lightweight MobileNetv2 network, thereby greatly reducing the computational parameters of the network while ensuring the ability to extract features. Secondly, in order to further reduce the number of model parameters and computational complexity, and minimize potential accuracy loss as much as possible, an improved depthwise separable convolution (DSC-SE-HsId) in this article was studied, which replaced some ordinary convolutions in the Neck and Head networks. Finally, in order to compensate for the partial accuracy loss caused by lightweight operations, and also to fuse feature maps of different scales to obtain more complete feature information, a new lightweight adaptive spatial feature fusion module (LSE-ASFF) in this article was studied and embedded behind the existing path aggregation network (PANet). In order to verify the performance and widespread applicability of the improved model, we conducted tests using the self-built lithium battery surface defect dataset, and the steel surface defect dataset provided by Northeastern University. Moreover, in order to prevent the network training from over fitting due to too few defect samples in the self-built dataset, we also used image enhancement to expand the dataset. And the experiment results show that the improved model proposed in this article achieves the highest TOPSIS score in both experimental datasets. Among them, compared with YOLOv4 on the self-built dataset, our improved model not only increases mAP50 by 2.97%, reaching 97.83%, but also has model parameters of only 18.16% of the original model, floating point of operations of only 13.87% of the original model, model size of only 21.02% of the original model, and model training time shortened by 30.67% compared to the original model. At last, the effectiveness and superiority of the improved model in this article are demonstrated through example analysis and comparison.

Multi-scale context UNet-like network with redesigned skip connections for medical image segmentation

Background and ObjectiveMedical image segmentation has garnered significant research attention in the neural network community as a fundamental requirement for developing intelligent medical assistant systems. A series of UNet-like networks with an encoder-decoder architecture have achieved remarkable success in medical image segmentation. Among these networks, UNet2+ (UNet++) and UNet3+ (UNet+++) have introduced redesigned skip connections, dense skip connections, and full-scale skip connections, respectively, surpassing the performance of the original UNet. However, UNet2+ lacks comprehensive information obtained from the entire scale, which hampers its ability to learn organ placement and boundaries. Similarly, due to the limited number of neurons in its structure, UNet3+ fails to effectively segment small objects when trained with a small number of samples. MethodIn this study, we propose UNet_sharp (UNet#), a novel network topology named after the “#” symbol, which combines dense skip connections and full-scale skip connections. In the decoder sub-network, UNet# can effectively integrate feature maps of different scales and capture fine-grained features and coarse-grained semantics from the entire scale. This approach enhances the understanding of organ and lesion positions and enables accurate boundary segmentation. We employ deep supervision for model pruning to accelerate testing and enable mobile device deployment. Additionally, we construct two classification-guided modules to reduce false positives and improve segmentation accuracy. ResultsCompared to current UNet-like networks, our proposed method achieves the highest Intersection over Union (IoU) values ((92.67±0.96)%, (92.38±1.29)%, (95.36±1.22)%, (74.01±2.03)%) and F1 scores ((91.64±1.86)%, (95.70±2.16)%, (97.34±2.76)%, (84.77±2.65)%) on the semantic segmentation tasks of nuclei, brain tumors, liver, and lung nodules, respectively. ConclusionsThe experimental results demonstrate that the reconstructed skip connections in UNet successfully incorporate multi-scale contextual semantic information. Compared to most state-of-the-art medical image segmentation models, our proposed method more accurately locates organs and lesions and precisely segments boundaries.

Mapping Geodiversity at a National Scale: the Case Study of Italy

In order to assess the geodiversity of the Italian Peninsula, which covers approximately 300.000 km2, a semi-quantitative method based on the use of grids recording several indicators and indices was developed. The variety of geological, geomorphological, and pedological elements, characterizing the Italian territory, has been assessed with a two-step procedure. Firstly, the variety algorithm has been applied using grid cells with variable size, related to the spatial resolution of the input data, then the resulting variety values were averaged with a fixed cell size functional to the extent of the study area and the output scale of the geodiversity map. This procedure made it possible to preserve the spatial resolution of the input data (Digital Terrain Model, lithological and soil maps) providing as output a geodiversity map that faithfully reproduces the features of the Italian territory. In case of discrete data (rivers, lakes, glaciers, etc.), a procedure that assigns to each cell the maximum area or length values out of all its elements has been implemented. It made possible to preserve the hydrological elements that shape the landscape (e.g., the longest rivers, largest lakes, etc.) and represent important freshwater resource. An overview of the geographical distribution of geodiversity classes over the whole Italian territory has been elaborated. The resulting geodiversity map is a valuable tool for environmental planning, in particular for the identification of areas to be preserved, for the proper management of geo-resources and natural services.

Building a multi‐scale, collaborative, and time‐integrated digital crust: The next stage of the Macrostrat data system

AbstractMacrostrat is a platform for deep‐time geoscientific research that integrates stratigraphic columns and geologic maps into a digital description of the crust. The database and supporting software track crustal evolution and provide location‐based geological information to geoscience end users. Macrostrat houses multiple scales of mapping and stratigraphic data, from continent‐ and basin level summaries to single quadrangles and measured sections. Currently, Macrostrat's primary data holdings consist of regional stratigraphic columns with a spatial footprint weighted heavily to North America. While the data are of sufficient scale and resolution to generate insights about Earth evolution, increasing resolution and expanding spatial coverage will allow a new generation of scientific and interpretive uses. The next phase of Macrostrat's development will increase the detail and complexity of Macrostrat's multiscale data holdings, largely by engaging a wider range of geoscientists in entering stratigraphic data. To support broad collaboration, we are building new web‐based software to assemble and visualize regional stratigraphic sequences, refine multiple working age models, and compose regional records from measured stratigraphic sections. These tools will allow Macrostrat to draw on the expertise of a wide range of geoscience workers and grow a dataset with global relevance and a variety of end uses. New capabilities will pave the way to processes for submission, review, coordination, and assimilation of community‐contributed stratigraphic datasets. Digital compilation of geological maps and columns requires substantial effort, and well‐designed systems for distributing this work in the geoscience community will allow Macrostrat to build more adaptable and scientifically relevant products.

A Last-100-Meters User Acceptance Issue: Navigating Fully Autonomous Vehicles through on-Board Navigation Systems

As the navigation interface of fully autonomous vehicles (FAV) will be wholly dependent on On-board Navigation Systems (ONS), it is foreseeable that ONS will replace steering wheels as the primary control mode of future FAVs. Thus, the ease of setting precise destinations using ONS may become a key factor affecting FAV user acceptance. A survey (N = 200) comprising Navigation Anxiety Questionnaire and User Preference Questionnaire reveals the user’s acceptable measurements of distance error and execution time when navigating FAVs in six common situations. Subsequently, an in-vehicle simulation test (N = 24) found most participants in the simulation test performed worse than the user’s acceptable measurements obtained through the questionnaire survey, which indicated that the existing ONS failed to meet users’ expectations for the FAV navigation experience. Findings of the questionnaire survey and the simulation test reveal a severe “last-100 meters” issue when controlling FAVs by ONS, and that destination familiarity, gender difference, text input behavior, map scale, and map display mode should be seriously considered when designing ONS for FAV control.

HS-YOLO: Small Object Detection for Power Operation Scenarios

Object detection methods are commonly employed in power safety monitoring systems to detect violations in surveillance scenes. However, traditional object detection methods are ineffective for small objects that are similar to the background information in the power monitoring scene, which consequently affects the performance of violation behavior detection. This paper proposed a small object detection algorithm named HS-YOLO, based on High-Resolution Network (HRNet) and sub-pixel convolution. First, to fully extract the microfeature information of the object, a small object feature extraction backbone network is proposed based on the HRNet structure. The feature maps of different scales are processed by multiple parallel branches and fused with each other in the network. Then, to fully retain the effective features of small objects, the sub-pixel convolution module is incorporated as the upsampling operator in the feature fusion network. The low-resolution feature map is upsampled to a higher resolution by reorganizing pixel values and performing padding operations in this module. On our self-constructed power operation dataset, the HS-YOLO algorithm achieved a mAP of 87.2%, which is a 3.5% improvement compared to YOLOv5. Particularly, the dataset’s AP for detecting small objects such as cuffs, necklines, and safety belts is improved by 10.7%, 5.8%, and 4.4%, respectively. These results demonstrate the effectiveness of our proposed method in detecting small objects in power operation scenarios.

An advanced YOLOv3 method for small object detection

Small object detection has important application value in the fields of autonomous driving and drone scene analysis. As one of the most advanced object detection algorithms, YOLOv3 suffers some challenges when detecting small objects, such as the problem of detection failure of small objects and occluded objects. To solve these problems, an improved YOLOv3 algorithm for small object detection is proposed. In the proposed method, the dilated convolutions mish (DCM) module is introduced into the backbone network of YOLOv3 to improve the feature expression ability by fusing the feature maps of different receptive fields. In the neck network of YOLOv3, the convolutional block attention module (CBAM) and multi-scale fusion module are introduced to select the important information for small object detection in the shallow network, suppress the uncritical information, and use the fusion module to fuse the feature maps of different scales, so as to improve the detection accuracy of the algorithm. In addition, the Soft-NMS and Complete-IOU (ClOU) strategies are applied to candidate frame screening, which improves the accuracy of the algorithm for the detection of occluded objects. The experimental results on MS COCO2017, VOC2007, VOC2012 datasets and the ablation experiments on MS COCO2017 datasets demonstrate the effectiveness of the proposed method.The experimental results show that the proposed method achieves better accuracy in small object detection than the original YOLOv3 model.

Real-World Non-Homogeneous Haze Removal by Sliding Self-Attention Wavelet Network

In complex natural haze scenes, image haze removal still faces significant challenges in removing non-homogeneous and dense haze. The double complexity of haze distribution, on the one hand, is reflected in the interference of haze to the global image information, and on the other hand, it is reflected in the imbalance of image brightness and color caused by random haze distribution. In natural scenes with prominent edge and texture features, the above problems may cause severe degradation of image quality and performance of various tasks. Numerous studies on network learning show that the effect of haze removal is closely related to haze feature expression. Therefore, to improve the performance of dehazing, this paper proposes a sliding self-attention wavelet network. Specifically, we first design a sliding self-attention module to identify haze regions in images and capture rich haze-related feature information. Then, considering the uneven distribution of haze in images, discrete wavelet transform (DWT) and inverse transform (IDWT) are used for constructing a hierarchical encoder-decoder structure, which can fully use the multi-resolution characteristics of DWT, locally decompose feature maps of different scales, extract low and high-frequency information, and then gradually recover sharp edges and precise texture details from hazy images. Finally, to enable the proposed network to generate more realistic haze-free images on different complex haze scenes, we develop a DWT-based adversarial loss function to constrain the low and high-frequency components of generated images closer to the corresponding clear images. Experimental results on the relevant public benchmark datasets show that the proposed algorithm achieves favorable dehazing performance.

Integrating Place and Sense of Place into Geovisualizations of Hurricane Storm Surge Risk

Abstract Geovisualizations play a central role in communicating hurricane storm surge risks to the public by connecting information about the hazard to a place. Meanwhile, people connect to places through meaning, functions, and emotional bond, known as a sense of place. The mixed-method approach presented in this paper focuses on the intersection of sense of place, geovisualization, and risk communication. We explored place meaning, scale of place, and place attachment in the coastal communities in Georgia and South Carolina. We conducted cognitive mapping focus groups and developed a series of geovisualizations of storm surge risk with varying representations of place. We then investigated people’s ability to connect visual storm surge information to a place and understand their risk by testing these geovisualizations in a large population survey (n = 1442). We found that a 2D regional-scale map displayed together with a 3D abstract representation of a neighborhood was the most helpful in enabling people to relate to a place, quickly make sense of the information, and understand the risk. Our results showed that while the geovisualizations of storm surge risk can be effective generally, they were less effective in several important and vulnerable groups. We found substantial impacts of race, income, map-reading ability, place attachment, and scale of place on how people connected the storm surge risk shown in the visual to a place. These findings have implications for future research and for considering the way weather forecasters and emergency managers communicate storm surge information with diverse audiences using geovisualizations. Significance Statement Weather forecasters and emergency managers often use geovisualizations to communicate hurricane storm surge risks and threats to the public. Despite the important role that geovisualizations play, few studies have empirically investigated their effectiveness in hazardous weather risk communication. With the overarching goal of understanding how geovisualizations enable coastal residents to understand and respond to risk, we use an interdisciplinary approach to create new knowledge about the effectiveness of geovisualizations in storm surge risk communication. Our results show substantial impacts of sociodemographic factors across many of the measures that enable people to connect to a place through visualizations. These findings have implications for communicating geospatially varying risk to diverse audiences.

Vertical displacement monitoring technique using radar interferometry data

All phenomena and processes occurring on the Earth's surface are closely related. Earth is characterized by internal and external planetary geological processes, which, throughout the entire geological development of the Earth, lead to its change. The speed and scale of geological processes change in time and space due to climatic changes. Changes are divided into long-term and momentary ones, which cause catastrophic phenomena, including landslides. An essential component of geoecological research is monitoring landslide processes using data from remote sensing of the Earth. The possibility of remote geoecological monitoring of landslide processes using satellite radar interferometry has been investigated, tested and experimentally substantiated. The right bank of the Kaniv Reservoir, with many registered landslides, was chosen as the test site. The results of the activity of vertical displacements of landslides for the spring period from 2015 to 2023 were obtained. Nine test sites and five control, stable areas affected by active surface deformations were investigated using 45 Sentinel-1A images. Geoecological monitoring of the activation of landslide processes at a detailed level was carried out using Sentinel-1 satellite images, a digital terrain model (DEM), topographic maps of various scales, and geological maps of Quaternary and pre-Quaternary structures. The advantage of the study of landslide processes by remote methods is the ability to quickly, on large areas, with relatively high accuracy and minimal economic costs, solve the problems of environmental protection to ensure the sustainable development of the environment and society.

Penilaian Kesesuaian Lahan untuk Tanaman Jagung dan Kedelai di Lahan Rawa Lebak Desa Tanjung Sejaro Kecamatan Indralaya Kabupaten Ogan Ilir

This study aims to assess the suitability of land for corn and soybean plants. The research has been carried out in January to March 2020 in the swampy land of Tanjung Sejaro Village, Indralaya District, Ogan Ilir Regency. This research uses an intensive level survey method with an area of 5 ha. Observation and measurement were made directly to the field using a 1: 5000 scale map. The sample points consist of 10 points taken by the grid method, at each point representing 0.5 ha. Soil samples taken at a depth of 0-20 cm from the ground surface. The variables observed in the field were that the drainage class was rather poor, the effective depth was generally >90 cm, the topography was flat (<3%). The parameters observed in the laboratory were chemical properties in the form of very acidic soil pH in the range of 3.97-4.37, CEC classified as low with a range of 7.50-15.00 cmol/kg, Total nitrogen was very low with a range of 0.01-0.08 %, Phosphorus available was classified as very low with a range of 0.25-4.60 ppm, Potassium available was classified as very low with a range of 0.19-0.38 cmol/kg. The results showed the actual suitability for corn and soybean plants was N (not suitable) with several factors limiting nutrient retention and available nutrients. If liming and fertilizing inputs are given, the potential suitability for corn and soybean plants is S2.

GIS learning and college students’ acquisition and understanding of spatial concepts

ABSTRACT Empirical evidence is insufficient on the specific roles GIS learning plays in developing students’ understanding various spatial concepts. The present study aims to draw attention to common struggles of learning some spatial concepts in geography and offer directions for future research on GIS learning and the development of student spatial concept lexicon. Three types of technical terms – neutral, helpful associations, and hindering associations – by Glessmer and Brose (2014) provided an excellent framework to explain why some concepts are more or less difficult for students to grasp. Our findings suggest that learning map scale and map projection are relatively easy because they are neutral and introduced as a new vocabulary. Overlay and density are terms provoking helpful associations because they carry a similar meaning in everyday language, so most students had no problems understanding the meaning of these concepts. Spatial association seemed to elicit unhelpful association because everyday use of the term, association, is not precise enough to define spatial association, resulting in students confused with other terms. Future research must be done in the context of GIS requiring students to be able to not only know spatial concepts but clearly articulate these concepts regarding various GIS applications.

Assessment of soil total phosphorus storage in a complex topography along China's southeast coast based on multiple mapping scales

Soil phosphorus (P) plays a vital role in both ecological and agricultural ecosystems, where total P (TP) in soil serves as a crucial indicator of soil fertility and quality. Most of the studies covered in the literature employ a single or narrow range of soil databases, which largely overlooks the impact of utilizing multiple mapping scales in estimating soil TP, especially in hilly topographies. In this study, Fujian Province, a subtropical hilly region along China's southeast coast covered by a complex topographic environment, was taken as a case study. The influence of the mapping scale on soil TP storage (TPS) estimation was analyzed using six digital soil databases that were derived from 3 082 unique soil profiles at different mapping scales, i.e., 1:50 000 (S5), 1:200 000 (S20), 1:500 000 (S50), 1:1 000 000 (S100), 1:4 000 000 (S400), and 1:10 000 000 (S1000). The regional TPS in the surface soil (0–20 cm) based on the S5, S20, S50, S100, S400, and S1000 soil maps was 20.72, 22.17, 23.06, 23.05, 22.04, and 23.48 Tg, respectively, and the corresponding TPS at 0–100 cm soil depth was 80.98, 80.71, 85.00, 84.03, 82.96, and 86.72 Tg, respectively. By comparing soil TPS in the S20 to S1000 maps to that in the S5 map, the relative deviations were 6.37%–13.32% for 0–20 cm and 0.33%–7.09% for 0–100 cm. Moreover, since the S20 map had the lowest relative deviation among different mapping scales as compared to S5, it could provide additional soil information and a richer soil environment than other smaller mapping scales. Our results also revealed that many uncertainties in soil TPS estimation originated from the lack of detailed soil information, i.e., representation and spatial variations among different soil types. From the time and labor perspectives, our work provides useful guidelines to identify the appropriate mapping scale for estimating regional soil TPS in areas like Fujian Province in subtropical China or other places with similar complex topographies. Moreover, it is of tremendous importance to accurately estimate soil TPS to ensure ecosystem stability and sustainable agricultural development, especially for regional decision-making and management of phosphate fertilizer application amounts.

Декомпозиция ретроспективных топографических карт для послойной оценки типов регионального природопользования

On the example of the most developed part of the lake Baikal basin, the decomposition of the digital coverage content from 12 sheets of the topographic map scale 1:84 000, compiled by the Corps of Military Topographers in 1896-1912, was performed. As a result of vectorization of retrospective layers of topographic objects and subsequent transformation of their shp files into classes of spatial objects of the ArcGIS environment, two cartographic sets of spatial object classes were formed that characterize two types of regional environmental management - traditional pastoral and invasive agricultural. As a result of layer-by-layer geoinformation modeling of cartographic sets on a modern topographic basis and a digital relief model, primary areas of settlement and economic use of the territory were established, physical and geographical characteristics of the past economic use of the territory were determined, and separate mapping of spatial and temporal life support structures of different types of nature management was performed. For operational digital visualization, a technological solution is proposed for storing and mapping sets of spatial object classes by organizing a file-based geodata of the ArcGIS information environment. Cartographic reconstructions of pastoral and agricultural nature management at the beginning of the 20th century are presented, as well as a general cartographic reconstruction of nature management on a landscape basis, which allows us to assert that traditional pastoral nature management does not involve the transformation of the Earth's surface and is characterized by continuous restorative dynamics of natural landscapes, and invasive nature management is based on the violation of soil cover, which later led to the emergence of areas of water and wind erosion.

Satellite-derived plant cover maps vary in performance depending on version and product

Understanding the accuracy and appropriate application scale of satellite-derived maps of vegetation cover is essential for effective management of the vast, remote rangelands of the world. However, the underlying models are updated frequently and may combine with rapidly changing vegetation conditions to cause variations in accuracy and precision over time. We sought to assess how model performance changed between different versions of satellite-derived cover products (Rangeland Analysis Platform, RAP, and Rangeland Condition Monitoring and Assessment Protocol, RCMAP) and how the performance of LandCart compared to RAP and RCMAP. Additionally, we asked how variability in agreement between LandCart and field-based models varied with scale. We utilized an intensive dataset of grid-point intercept functional group cover data collected between 2016 and 2020 across the ∼113 kHA 2015 Soda Wildfire to 1) evaluate r2 agreement between versions of each satellite-derived product and plot-level field data and 2) assess relative standard error of agreement in cover between LandCart and continuous field-based Empirical Bayesian Kriging (EBK) regression models. Agreement between satellite- compared to field-plot values of cover (r2) increased for RCMAP Version 5.0 compared to Version 2.0, but there were negligible changes between versions of RAP. Despite this, r2 values of RCMAP and LandCart were nearly always less than RAP. Variability in agreement between EBK regression model cover and LandCart-derived cover decreased with the scale of consideration. Variability in agreement between satellite-derived cover products and field-based metrics is lowest at larger scale (mega-fire or regional) and varies from year to year and across versions, which could complicate detection of temporal changes in plant cover.

Overview of large scale map production with UAV based photogrammetric technique: A case study in Izmir-Cesme territory of Turkey

UAVs, also known as unmanned aerial vehicles, have emerged as an efficient and flexible system for offering a rapid and cost-effective solution. In recent years, large-scale mapping using UAV photogrammetry has gained significant popularity and has been widely adopted in academia as well as the private sector. This study aims to investigate the technical aspects of this field, provide insights into the procedural steps involved, and present a case study conducted in Cesme, Izmir. The findings derived from the case study are thoroughly discussed, and the potential applications of UAV photogrammetry in large-scale mapping are examined. The study area is divided into 12 blocks. The flight plans and the distrubition of ground control point (GCP) locations were determined based on these blocks. As a result of the data processing procedure, average GCP positional errors ranging from 1 to 18 cm have been obtained for the blocks.

Geomorphology of the Central Kamchatka Depression, the Kamchatka Peninsula, NE Pacific

ABSTRACT The Kamchatka Peninsula lies on the eastern active margin of Eurasia, adjacent to the Kuril-Kamchatka subduction zone. In this study, we provide a geomorphological map of the Central Kamchatka Depression – the largest sedimentary basin in Kamchatka and also in all the island arcs of the North Pacific. The depression extends along Kamchatka at latitudes ∼44–46° N for more than 400 km and is up to 100 km wide. The geomorphological map of the Central Kamchatka Depression is designed at a scale of 1:500,000 and contains ∼3000 mapped objects. Volcanic and tectonic landforms have been collected from published sources and generalised to the map scale. Glacial and fluvial landforms have been mapped using remote sensing data and field geomorphological surveys. Areas of ambiguous interpretation have been outlined. This map is the most detailed geomorphological map published for this region that may guide further geomorphological and paleogeographical investigations.

Convolutional Neural Network Defect Detection Algorithm for Wire Bonding X-ray Images.

To address the challenges of complex backgrounds, small defect sizes, and diverse defect types in defect detection of wire bonding X-ray images, this paper proposes a convolutional-neural-network-based defect detection method called YOLO-CSS. This method designs a novel feature extraction network that effectively captures semantic features from different gradient information. It utilizes a self-adaptive weighted multi-scale feature fusion module called SMA which adaptively weights the contribution of detection results based on different scales of feature maps. Simultaneously, skip connections are employed at the bottleneck of the network to ensure the integrity of feature information. Experimental results demonstrate that on the wire bonding X-ray defect image dataset, the proposed algorithm achieves mAP 0.5 and mAP 0.5-0.95 values of 97.3% and 72.1%, respectively, surpassing the YOLO series algorithms. It also exhibits certain advantages in terms of model size and detection speed, effectively balancing detection accuracy and speed.

Invited Session IV: Studies of the visual cortex with sub-millimeter resolution: Toward an all-optical bi-directional interrogation of topographic population codes in primate cortex.

The representation of visual stimuli in primate V1 is widely distributed and topographic. This raises the possibility that in some visual tasks, downstream areas that decode V1 signals in order to mediate perception could combine V1 signals at a relevant topographic scale-e.g., at the scale of orientation columns. If this were the case, then the fundamental unit of information would be individual columns rather than single neurons, and to account for the subject's behavior in a perceptual task, it would be necessary and sufficient to consider the summed activity of the thousands of neurons within each column. In this presentation I will discuss our initial attempts to test this topographic-code hypothesis using our optical-genetic toolbox for "reading" and "writing" neural population codes at the spatial scales of topographic maps in V1 of behaving macaques.