Find research datasets worth reusing

Incorporating Onboard Human Detection into Cyborg Insect

Eight features (VAR, MAD, CoV, AAC, WA, LCoV, LDASDV, WL) are extracted from thermal images using a microcontroller and sent wirelessly to a remote PC for data collection. The data collection includes both "human" and "no human" classes. The temperature during data collection is kept between 24°C and 26.5°C. To collect human class data features, a person sits in front of the wireless backpack at distances ranging from 25 cm to 1 meter. The collected feature datasets  (12,062 x 8) are split into 70% for training and 30% for testing. The RF classifier is implemented with ten trees to reduce memory usage, resulting in a compact size of 1.65 KB. This embedded Random Forest (RF) classifier can efficiently operate on the proposed backpack with its 32KB SRAM and 256KB flash memory. The average processing time from feature extraction to the random forest algorithm for onboard human detection is 7.5 ms.

Supplementary Material for "Detecting Automatic Software Plagiarism via Token Sequence Normalization"

This repository contains additional material supporting the paper titled "Detecting Automatic Software Plagiarism via Token Sequence Normalization", presented at ICSE 2024 (research track). <div> <div> <div> The paper presents a defense mechanism against automated plagiarism generators utilizing program dependence graphs and demonstrates its effectiveness in countering insertion-based and reordering-based obfuscation attacks. </div> </div> </div> The defense mechanism was also integrated into the software plagiarism detector <a title="JPlag Repository on GitHub" href="https://github.com/jplag/JPlag">JPlag</a>, thus providing a widely accessible solution. Contents Overview: <ul> <li> Datasets: Two datasets from the <a title="PROGpedia Repository" href="../record/7449056">PROGpedia</a> collection and two internal datasets. For the latter, only the metadata is available due to the sensitive nature of the data. </li> <li> Plagiarized Submissions: Generated plagiarism instances illustrating various obfuscation methods such as insertion, reordering, and insert-after-reordering. </li> <li> Evaluation Data: JSON files detailing calculated similarities and runtime measurements for all datasets. </li> <li> Source Code: The implementation of our defense mechanism based on the software plagiarism detector <a title="JPlag Repository on GitHub" href="https://github.com/jplag/JPlag">JPlag</a> (v4.0.0). Note that JPlag is licensed under the GPL-3.0 license. </li> <li> Evaluation Code: Python code for runtime measurements. </li> <li> Interactive Plots: HTML visualizations of the paper's plots, offering dynamic insights into the research findings. particularly focusing on the detection of automatic software plagiarism through token sequence normalization. </li> <li>Demo: A packaged JAR of our implementation alongside an instruction on how to execute it.</li> </ul>

Figure S1. Determination of the limit of detection (LOD) of S2 subunit.

Open the record for dataset details and reuse information.

opencc-by-4.0Nov 2023View details →

Supplementary data for "Validation of a quantitative PCR assay for detection of sugar kelp (Saccharina latissima) in coastal marine sediments"

Raw data for a paper in preparation: "Validation of a quantitative PCR assay for detection of sugar kelp (Saccharina latissima) in coastal marine sediments".Contains:<ol><li>Spreadsheet of curated raw data used to calculated S. latissima COI gene copy number per gram of dry sediment for a biomass spiking experiment to determine how gene copy number correlates with known amounts of added kelp biomass.</li><li>Spreadsheet of curated raw data from a field study looking at S. latissima gene copy numbers in field samples from the Damariscotta River estuary in Maine, from three sites, one under a small kelp farm, one under a site with a single kelp line, and another in a control site with no known influence of kelp farming activity.</li></ol>

dryad32/100

A novel approach to low-cost, rapid and simultaneous colorimetric detection of multiple analytes using 3D printed microfluidic channels

This research paper presents an inventive technique to swiftly create microfluidic channels on distinct membrane papers, enabling colorimetric drug detection. Using a modified DIY RepRap 3D printer with a syringe pump, microfluidic channels (µPADs) are crafted on a flexible nylon-based substrate. This allows simultaneous detection of four common drugs with a single reagent. An optimized blend of Polydimethylsiloxane (PDMS) dissolved in hexane is used to create hydrophobic channels on various filter papers. The PDMS-hexane mixture infiltrates the paper's pores, forming hydrophobic barriers that confine liquids within the channels. These barriers are cured on the printer's hot plate, controlling channel width and preventing spreading. Capillary action drives fluid along these paths without spreading. This novel approach provides a versatile solution for rapid microfluidic channel creation on membrane papers. The DIY RepRap 3D printer integration offers precise control and faster curing. The PDMS-hexane solution accurately forms hydrophobic barriers, containing liquids within desired channels. The resulting microfluidic system holds potential for portable, cost-effective drug detection and various sensing applications.

opencc-zeroJan 2024View details →

Supplementary Material for "Automated Detection of AI-Obfuscated Plagiarism in Modeling Assignments"

This repository contains additional material supporting the paper titled "Automated Detection of AI-Obfuscated Plagiarism in Modeling Assignments", presented at ICSE 2024 (SEET track). The paper presents a token-based approach for detecting modeling plagiarism. It leverages a novel normalization technique to achieve resilience against common obfuscation attacks. The approach was also integrated into the software plagiarism detector <a title="JPlag Repository on GitHub" href="https://github.com/jplag/JPlag">JPlag</a>, thus providing a widely accessible solution. Contents Overview: <ul> <li>Source Code: The implementation of our approach (contribution 1) based on the software plagiarism detector <a title="JPlag Repository on GitHub" href="https://github.com/jplag/JPlag">JPlag</a> (v4.0.0). Note that JPlag is licensed under the GPL-3.0 license.</li> <li>ChatGPT Study: An exploration of how ChatGPT can be exploited for cheating in modeling assignments.</li> <li>Datasets: The three datasets of our evaluation based on EMF modeling assignments.</li> <li>Raw Results: All raw data of our evaluation results, as used in our plots and tables.</li> <li>Demo: A packaged JAR of our approach's implementation alongside an instruction on how to use it.</li> </ul>

TinyWT: A Large-Scale Wind Turbine Dataset of Satellite Images for Tiny Object Detection

This dataset is from the paper "TinyWT: A Large-Scale Wind Turbine Dataset of Satellite Images for Tiny Object Detection", which has been accepted by the WACV 2024 CV4EO Workshop.

opencc-by-nc-sa-4.0Oct 2023View details →

Dataset for "Enhancing Cloud Detection in Sentinel-2 Imagery: A Spatial-Temporal Approach and Dataset"

This dataset is built for time-series Sentinel-2 cloud detection and stored in Tensorflow TFRecord (refer to https://www.tensorflow.org/tutorials/load_data/tfrecord). Each file is compressed in 7z format and can be decompressed using Bandzip or 7-zip software. Dataset Structure: Each filename can be split into three parts using underscores. The first part indicates whether it is designated for training or validation ('train' or 'val'); the second part indicates the Sentinel-2 tile name, and the last part indicates the number of samples in this file. For each sample, it includes: <ol> <li>Sample ID;</li> <li>Array of time series 4 band image patches in 10m resolution, shaped as (n_timestamps, 4, 42, 42);</li> <li>Label list indicating cloud cover status for the center $6\times6$ pixels of each timestamp;</li> <li>Ordinal list for each timestamp;</li> <li>Sample weight list (reserved);</li> </ol> Here is a demonstration function for parsing the TFRecord file: <pre><code>import tensorflow as tf # init Tensorflow Dataset from file name def parseRecordDirect(fname): sep = '/' parts = tf.strings.split(fname,sep) tn = tf.strings.split(parts[-1],sep='_')[-2] nn = tf.strings.to_number(tf.strings.split(parts[-1],sep='_')[-1],tf.dtypes.int64) t = tf.data.Dataset.from_tensors(tn).repeat().take(nn) t1 = tf.data.TFRecordDataset(fname) ds = tf.data.Dataset.zip((t, t1)) return ds keys_to_features_direct = { 'localid': tf.io.FixedLenFeature([], tf.int64, -1), 'image_raw_ldseries': tf.io.FixedLenFeature((), tf.string, ''), 'labels': tf.io.FixedLenFeature((), tf.string, ''), 'dates': tf.io.FixedLenFeature((), tf.string, ''), 'weights': tf.io.FixedLenFeature((), tf.string, '') } # The Decoder (Optional) class SeriesClassificationDirectDecorder(decoder.Decoder): """A tf.Example decoder for tfds classification datasets.""" def __init__(self) -> None: super().__init__() def decode(self, tid, ds): parsed = tf.io.parse_single_example(ds, keys_to_features_direct) encoded = parsed['image_raw_ldseries'] labels_encoded = parsed['labels'] decoded = tf.io.decode_raw(encoded, tf.uint16) label = tf.io.decode_raw(labels_encoded, tf.int8) dates = tf.io.decode_raw(parsed['dates'], tf.int64) weight = tf.io.decode_raw(parsed['weights'], tf.float32) decoded = tf.reshape(decoded,[-1,4,42,42]) sample_dict = { 'tid': tid, # tile ID 'dates': dates, # Date list 'localid': parsed['localid'], # sample ID 'imgs': decoded, # image array 'labels': label, # label list 'weights': weight } return sample_dict # simple function def preprocessDirect(tid, record): parsed = tf.io.parse_single_example(record, keys_to_features_direct) encoded = parsed['image_raw_ldseries'] labels_encoded = parsed['labels'] decoded = tf.io.decode_raw(encoded, tf.uint16) label = tf.io.decode_raw(labels_encoded, tf.int8) dates = tf.io.decode_raw(parsed['dates'], tf.int64) weight = tf.io.decode_raw(parsed['weights'], tf.float32) decoded = tf.reshape(decoded,[-1,4,42,42]) return tid, dates, parsed['localid'], decoded, label, weight t1 = parseRecordDirect('filename here') dataset = t1.map(preprocessDirect, num_parallel_calls=tf.data.experimental.AUTOTUNE) # </code></pre> Class Definition: <ul> <li>0: clear</li> <li>1: opaque cloud</li> <li>2: thin cloud</li> <li>3: haze</li> <li>4: cloud shadow</li> <li>5: snow</li> </ul> Dataset Construction: First, we randomly generate 500 points for each tile, and all these points are aligned to the pixel grid center of the subdatasets in 60m resolution (eg. B10) for consistence when comparing with other products.  It is because that other cloud detection method may use the cirrus band as features, which is in 60m resolution.  Then, the time series image patches of two shapes are cropped with each point as the center. The patches of shape $42 \times 42$ are cropped from the bands in 10m resolution (B2, B3, B4, B8) and are used to construct this dataset. And the patches of shape $348 \times 348$ are cropped from the True Colour Image (TCI, details see sentinel-2 user guide) file and are used to interpreting class labels. The samples with a large number of timestamps could be time-consuming in the IO stage, thus the time series patches are divided into different groups with timestamps not exceeding 100 for every group.

MOTS-annotated UAV Vineyard Dataset captured using Multiple Perspectives to avoid Leaf Occlusion for Object Detection and Tracking

This dataset contains UAV RGB videos (MP4) recorded with a Phantom4 RTK in a vineyard during the harvesting campaign of 2023. It also includes frames and annotations (PNG) to boost Object Detection and Tracking of grape bunches. There are two types of videos: (1) videos capturing the side of the canopy from a frontal point of view only, and (2) videos that collect the data from multiple perspectives to avoid leaf occlusion, common in commercial vineyards. All flights were executed 3 meters above ground level, with a clear sky and wind speed below 0.5 m/s.

Detecting Usage of Deprecated Web APIs via Tracing: Replication Package

Replication package for "Detecting Usage of Deprecated Web APIs via Tracing", published at 21st IEEE International Conference on Software Architecture (ICSA 2024)   Changes from v2 to v3: <ul> <li>readme.md: The requirements are now stated and explanations are given in more detail. Some typos in commands are fixed.</li> </ul>   Changes from v1 to v2: <ul> <li>DATA: In addition to the intermedia data (output.json), the overall result files (results.json) are now provided as well for all example projects.</li> <li>IMPLEMENTATION: An incomplete version of the DeprecationDetector was present in v1. The new version includes extended matching functionality for specifications and improved export formatting.</li> <li>PROJECT: The server.jar in the client-server example project has been improved.</li> </ul>

opencc-by-4.0Oct 2023View details →

A detectable change in the air-sea CO2 flux estimate from sailboat measurements

This repository contains files and scripts used to generate figures for the paper titled "A detectable change in the air-sea CO2 flux estimate from sailboat measurements". The study quantifies the impact of pCO2 observations measured by the sailboat "Seaexplorer" on the air-sea CO2 flux estimate. Overview:We used pCO2 measurements sourced from SOCAT (www.socat.info), along with other environmental variables. We applied the 2-step neural network method SOM-FFN (Landschützer et al., 2013) to reconstruct air-sea CO2 flux estimates.Four sets of air-sea CO2 flux estimates were reconstructed, each with 40 ensemble members to account for random errors in the neural network. These are the four sets:E1) fluxesA - Based on SOCATv2022 including Seaexplorer dataE2) fluxesB - Based on SOCATv2022 excluding Seaexplorer data.E3) fluxesC - Based on SOCATv2022 data, including Seaexplorer data, but introducing a random uncertainty of ± 5 μatm to the Seaexplorer data.E4) fluxesD - Based on SOCATv2022 data, including Seaexplorer data, but introducing a constant measurement offset of 5 μatm to the Seaexplorer data.These air-sea CO2 flux density estimates are 4-dimensional, represented in mol C m-2 yr-1, with the following dimensions:- Ensemble size of flux reconstructions (40)- Time (months from 1982 to 2021, totaling 480 months)- Latitude (1-degree intervals, totaling 180 latitudes)- Longitude (1-degree intervals, totaling 360 longitudes)

opencc-by-4.0Feb 2024View details →

Crowd Detection Dataset

<h1>Growd Detection Dataset</h1> This dataset comprises aerial images captured from varying altitudes using UAVs. The dataset includes annotations in the form of both bounding boxes and point annotations, totaling 84,486 annotations specifically identifying human heads. <h2> Dataset Details</h2> <ul> <li>AgiosGeorgios_27.5 & Civil Defense <ul> <li>Photos of civil defense individuals (crowd of ~10 & ~20) from high and medimum altitudes, both far and near the sea shore </li> </ul> </li> </ul>   <ul> <li>Kornos_labeling <ul> <li>\DJI_0042 <ul> <li>Photos of individuals (crowd of ~15) from high and medimum altitudes</li> </ul> </li> <li>\DJI_0049 <ul> <li>Photos of individuals (crowd of ~30) far away and near</li> </ul> </li> <li>\DJI_0050 <ul> <li>Photos of individuals (crowd of ~30) from high and low altitutes</li> </ul> </li> <li>\DJI_0085 <ul> <li>Photos of individuals (crowd of ~30) from far away and near</li> </ul> </li> </ul> </li> </ul>   <ul> <li>UCY Campus - Low altitude <ul> <li>\data <ul> <li>Photos of individuals (crowd of ~10) from low altitutes at the UCY campus</li> </ul> </li> </ul> </li> </ul>   <ul> <li>UCY Campus <ul> <li>\part1 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus</li> </ul> </li> <li>\part2 <ul> <li>Photos of individuals from high altitude in campus, in a parking area at UCY campus</li> </ul> </li> <li>\part3 <ul> <li>Photos of individuals from high altitude in campus, in a parking area at UCY campus</li> </ul> </li> <li>\part4 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus</li> </ul> </li> <li>\part5 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus</li> </ul> </li> </ul> </li> </ul>   <ul> <li>UCY_2023 <ul> <li>\part1 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus</li> </ul> </li> <li>\part2 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus (not annotated, can be used for testing)</li> </ul> </li> <li>\part3 <ul> <li>Photos of individuals from high altitude in campus, mostly near of parking area at UCY campus (not annotated, can be used for testing)</li> </ul> </li> </ul> </li> </ul>

Apple detection dataset

1.背景介绍 苹果检测是计算机视觉与图像处理领域中的一个重要问题,其应用包括农业自动化、食品质量检测和无人机水果采摘等。在这些应用中,准确地检测和识别苹果对于提高生产效率、保证食品质量以及优化农业管理都具有重要意义。苹果检测的挑战之一在于其外观和形状多样,受光照、阴影、颜色变化以及遮挡等因素的影响较大。另外,苹果通常生长在树上,背景复杂,可能包含树叶、树枝、其他水果、土壤等干扰物体,因此需要建立鲁棒的检测方法来应对这些复杂情况。近年来,深度学习技术的发展为苹果检测带来了新的希望。 基于深度学习的目标检测算法,如YOLO (You Only Look Once)和Faster R-CNN等,在苹果检测中取得了显著的成果。这些算法能够针对复杂背景和不同形状的苹果进行准确的检测和定位,为后续的质量评估和自动化采摘提供了可靠的基础。除了算法技术的发展,苹果检测还涉及到大规模数据集的采集和标注工作。精心采集的数据集可以帮助算法更好地理解苹果的外观特征,并提高检测的准确性和鲁棒性。同时,合适的数据增强技术也能够有效地提升算法在复杂场景下的表现。 总的来说,苹果检测是一个具有挑战性但又充满潜力的研究领域。随着计算机视觉和深度学习技术的不断进步,相信苹果检测技术将在农业生产和食品加工领域发挥越来越重要的作用,为农业现代化和智能化注入新的活力。 2.数据集介绍 本研究采用LabellImg标注工具对原始图像中的苹果目标绘制外接矩形框以实现人工标注。设定苹果标签为apple,保存格式为PASCAL VOC。苹果图像数据集共2825张。 2.1数据格式 苹果检测数据集包括训练集、验证集、测试集和苹果的标注数据。苹果图像数据集共2825张。将数据集按照9:1:1的比例随机分为2197幅图像用来训练模型, 314幅图像用来验证模型,314幅图像用来测试模型。 2.2文件列表 本数据集包含如下文件: 1. Train:训练集 2. Valid:验证集 3. Test:测试集 4.  Data.yaml: 描述了数据集的相关信息,包括数据文件的路径、标签信息、数据格式 注:使用解压工具解压可能会失败,请使用命令行进行解压操作,如:unzip xxx.zip 3.数据集类型 制造、图像识别

opencc-by-4.0Sep 2023View details →

Data associated with global-bird-detection-paper

Data for repository located at https://github.com/sarabsethi/global-bird-detection-paper Unzip the file to the directory and follow the instructions in the repository to run the code. (March 2024 update reflects modifications made for manuscript revision)

Apple detection code

1.背景介绍 苹果检测是计算机视觉与图像处理领域中的一个重要问题,其应用包括农业自动化、食品质量检测和无人机水果采摘等。在这些应用中,准确地检测和识别苹果对于提高生产效率、保证食品质量以及优化农业管理都具有重要意义。苹果检测的挑战之一在于其外观和形状多样,受光照、阴影、颜色变化以及遮挡等因素的影响较大。另外,苹果通常生长在树上,背景复杂,可能包含树叶、树枝、其他水果、土壤等干扰物体,因此需要建立鲁棒的检测方法来应对这些复杂情况。近年来,深度学习技术的发展为苹果检测带来了新的希望。   基于深度学习的目标检测算法,如YOLO (You Only Look Once)和Faster R-CNN等,在苹果检测中取得了显著的成果。这些算法能够针对复杂背景和不同形状的苹果进行准确的检测和定位,为后续的质量评估和自动化采摘提供了可靠的基础。除了算法技术的发展,苹果检测还涉及到大规模数据集的采集和标注工作。精心采集的数据集可以帮助算法更好地理解苹果的外观特征,并提高检测的准确性和鲁棒性。同时,合适的数据增强技术也能够有效地提升算法在复杂场景下的表现。   总的来说,苹果检测是一个具有挑战性但又充满潜力的研究领域。随着计算机视觉和深度学习技术的不断进步,相信苹果检测技术将在农业生产和食品加工领域发挥越来越重要的作用,为农业现代化和智能化注入新的活力。 2.代码介绍 本研究采用Python语言对YOLOv8s苹果检测系统进行设计改进。 2.1文件列表 1.github ISSUE_TEMPLATE:提供不同类型的问题报告模板,包括 bug-report,yml、config,yml!feature-requestym和 question.yml。这些模板帮助用户以结构化的方式报告错误、提出功能请求或提问。 workfows:包含多个工作流文件,如ciym(持续集成)、cla.yml(贡献者许可协议)、codeqlym(代码质量检查)、 docker.ym(Docker配置)、greetngs.yml(自动问候新贡献者)、links,ymlkpublish.yml(自动发布)、stale.yml(处理陈旧问题) dependabot:yml(自动依赖更新) 这些文件共同支持项目的自动化管理,包括代码质量保证、持续集成和部署、社区互动和依赖项维护。 2.docker Dockerfile:主要的Docker配置文件,用于构建项目的默认Docker镜像。 Dockerfile-arm64:针对ARM64架构的设备(如某些类型的服务器或高级嵌入式设备)定制的Docker配置, Dockerfile-conda:使用Conda包管理器配置环境的Docker配置文件: Dockerfile-cpu: 为不支持GPU加速的环境配置的Docker配置文件。 Dockerfile-jetson: 专为NVIDlA Jetson平台定制的Docker配置。 Dockerfile-python: 可能是针对纯Python环境的简化Docker配置 Dockerfile-runner 可能用于配置持续集成特续部署(CI/CD)运行环境的Docker配置, 这些配置文件是用来部署用的,用户可以根据自己的需要选择合适的环境来部署和运行项目 3.docs docs日录通常用于存放文档资料,包括多种语言的翻译。例如,此目录下有多个文件夹,每个文件夹代表一种语言(如en代表英语文档)。除此之外,还有几个重要的Python脚本和配置文件给大家说一下: build_docs.py:-个Python脚本,用于自动化构建和编译文档的过程,mkdocs.yml:MkDocs配置文件,用于指定文档网站的结构和设置。 以mkdocs_es.yml为例,这是用于构建西班牙语文档的MKDocs配置文件。类似的,mkdocs zh.yml用于构建中文文档。 4.examples YOLO8-CPP-nference:包含C++语言实现的YOLOV8推理示例,内有CMakeLists.bt(用于项目构建的CMake配置文件)inference.cpp和inferenceh(推理相关的源代码和头文件),main.cpp(主程序入口)以及README.md(使用说明) YOLOv8-0NNXRuntime:提供Pvthon语言与ONNXRunime结合使用的YOLOV8推理示例,其中main.py是主要的脚本文件README.md提供了如何使用该示例的指南。 YOLOY8-0NNXRuntime-CPP:与上述ONNXRuntime类似,但是是用C++编写的,包含了相应的CMakeLists bt,inference.cpp. inference.h和main.cpp文件,以及用于解释如何运行示例的README.md。 每个示例都配有相应的文档,是当我们进行模型部署的时候在不同环境中部署和使用YOLOv8的示例. 5.tests conftest.py:包含测试配置选项或共享的测试助手函数test_cli.py:用于测试命令行界面(CLI)的功能和行为。test_cuda.py:专门测试项目是否能正确使用NVIDIA的CUDA技术,确保GPU加速功能正常test_engine.py:测试底层推理引擎,如模型加载和数据处理等。test integrations.py:测试项目与其他服务或库的集成是否正常工作。test_python.py:用于测试项目的Python AP|接口是否按预期工作。 6.runs:训练结果 7.utralytics datasets文件夹:包含数据集的配置文件,如数据路径、类别信息等(就是我们训练YOLO模型的时候需要一个数据集,这里面就保存部分数据集的yaml文件,如果我们训练的时候没有指定数据集则会自动下载其中的数据集文件,但是很容易失败!)。models文件夹:存放模型配置文件,定义了模型结构和训练参数等,这个是我们改进或者就基础版本的一个yam文件配置的地方。 nodelsx件天中的每yamlz件代表了不同的YOLOv8模型配置,具体包括: yolov8.yaml:这是YOLOv8模型的标准配置文件,定义了模型的基础架构和参数。 yolov8-cls.yam!: 配置文件调整了YOLOv8模型,专门用于图像分类任务。 yolov8-ghost.yaml: 应用Ghost模块的YOLOv8变体,旨在提高计算效率:yolov8-ghost-p2.yaml和 yolov8-ghost-p6.yaml: 这些文件是针对特定大小输入的Ghost模型变体配置。yolov8-p2.yaml和 yolov8-p6.yam: 针对不同处理级别(例如不同的输入分辨率或模型深度)的YOLOv8模型配置 yolov8-pose.yaml: 为姿态估计任务定制的YOLOv8模型配置。 yolov8-pose-p6.yam!: 针对更大的输入分辨率或更复杂的模型架构姿态估计任务。 yolov8-rtdetr.yaml: 可能表示实时检测和跟踪的YOLOv8模型变体。 yolov8-seg.yaml和 yolov8-seg-p6.yaml: 这些是为语义分割任务定制的YOLOv8模型配置 trackers文件夹:用于追踪算法的配置。 init .py文件:表明`cfg是-个Python包。 default.yaml:项目的默认配置文件,包含了被多个模块共享的通用配置项。 8.data download weights.sh:用来下载预训练权重的脚本。get_coco.sh,get_coco128.sh,get imagenet.sh:用于下载COCO数据集完整版、128张图片版以及lmageNet数据集的脚本 在data文件夹中,包括: annotator.py:用于数据注释的工具。 augment.py: 数据增强相关的函数或工具。base.py,build.py,converter.py: 包含数据处理的基础类或函数、构建数据集的脚本以及数据格式转换工具 dataset.py:数据集加载和处理的相关功能。 loaders.py: 定义加载数据的方法。 utils.py:各种数据处理相关的通用工具函数。 9.engine exporter.py:用于将训练好的模型导出到其他格式,例如ONNX或TensorRT。 model.py: 包含模型定义,还包括模型初始化和加载的方法。 predictor.py:包含推理和预测的逻辑,如加载模型并对输入数据进行预测。 results.py:用于存储和处理模型输出的结果。 trainer.py:包含模型训练过程的逻辑。 tuner.py: 用于模型超参数调优。 validator.py: 包含模型验证的逻辑,如在验证集上评估模型性能。 10.models classify:这个目录可能包含用于图像分类的YOLO模型。 detect: 包含用于物体检测的YOLO模型 pose:包含用于姿态估计任务的YOLO模型 segment: 包含用于图像分割的YOLO模型 11.nn modules双件夹: init .py: 表明此目录是Python包。 block.py: 包含定义神经网络中的基础块,如残差块或瓶颈块。 conv.py: 包含卷积层Q 相关的实现。 head.py: 定义网络的头部,用于预测。 transformer.py:包含Transformer模型相关的实现 utils.py: 提供构建神经网络时可能用到的辅助函数, init .py:同样标记这个目录为Python包 autobackend.py: 用于自动选择最优的计算后端, tasks,py,定义了使用神经网络完,成的不同任务的流程,例如分类、检测或分割,所有的流程基本上都定义在这里,定义模型前向传播都在这里。 12.solutions init_.py: 标识这是-个Python包。 ai_gym.py:与强化学习相关,例如在OpenAlGym环境中训练模型的代码,heatmap.py:用于生成和处理热图数据,这在物体检测和事件定位中很常见。object counter.py: 用于物体计数的脚本,包含从图像中检测和计数实例的逻辑。 13.utils callbacks.py:包含在训练过程中被调用的回调函数。autobatch.py:用于实现批处理优化,以提高训练或推理的效率:benchmarks.py: 包含性能基准测试相关的函数.checks.py 用于项目中的各种检查,如参数验证或环境检查。dist.py:涉及分布式计算相关的工具 downloads.py:包含下载数据或模型等资源的脚本,errors.py:定义错误处理相关的类和函数, fles.py: 包含文件操作相关的工具函数。 instance.py: 包含实例化对象或模型的工具, loss.py: 定义损失函数, metrics.py: 包含评估模型性能的指标计算函数。 ops.py: 包含自定义操作,如特殊的数学运算或数据转换patches.py:用于实现修改或补丁应用的工具plotting.py: 包含数据可视化相关的绘图工具。 tal.py:一些损失函数的功能应用 torch utils.py:提供PyTorch相关的工具和辅助函数,包括GFLOPs的计算triton.py: 可能与NVIDlA Triton Inference Server集成相关tuner.py: 包含模型或算法调优相关的工具。 3.模型模型 本研究模型主要在YOLOv8s主干,颈部和检测头部分进行改进,将MobileNetV3替换原始主干网络,MobileNetV31.结合了硬件感知的网络架构搜索(NAS)和NetAdapt算法,针对移动设备CPU进行优化,引入了新颖的架构设计,包括反转残差结构和线性瓶颈层。提出了高效的Lite Reduced Atrous $patial Pyramid Pooling(LR-ASPP)作为新的分割解码。 在颈部部分引入BiFF双向金字塔网络,BiFF有高效的双向跨尺度连接:BIFPN通过在自顶向下和自底向上路径之间建立双向连接,允许不同尺度特征间的信息更有效地流动和融合。简化的网络结构:BIFPN通过删除只有一个输入边的节点、在同一层级的输入和输出节点间添加额外边,以及将每个双向路径视为-特征网络层并重复多次,来优化跨尺度连接。加权特征融合:BIFPN引入了可学习的权重来确定不同输入特征的重要性,从而提高了特征融合的效果。 在检测头部分引入ASFF思想:自适应空间特征融合:提出了一种新的金字塔特征融合策略,能够空间过滤,中突信息,压制不同尺度特征间的不一致性。改善尺度不变性:通过ASFF策略,显著提升了特征的尺度不变性,有助于提高对象检测的准确性。低推理开销:在提升检测性能的同时,几乎不增加额外的推理开销。

opencc-zeroMar 2024View details →

dryad32/100

SiMPL Wildlife Magnets: A camera trap tool for detecting all creatures great and small

Small mammals compose a substantial portion of the seed predators and dispersers, as well as the prey, in many ecosystems. Nevertheless, information on distribution and habitat use is limited, partly because inexpensive, minimally -invasive surveying of small mammals has remained challenging. We created the SiMPL wildlife magnet – a baited camera trap design that allows passive monitoring of wildlife, especially small- to medium-sized mammals. The SiMPL magnet is inexpensive and easy to construct. To evaluate its effectiveness, we conducted a case study using 10 sites operating for two years along an elevation gradient in the White Mountains of the northeastern U.S. We found that the SiMPL wildlife magnet can be used to detect rodents, mesocarnivores, and, to a lesser extent, ungulates. We examined how the detection probability of mammal species changed with the inclusion of a SiMPL wildlife magnet using data from pre- and post-establishment. We found a significant increase in community-level detection probability with the use of SiMPL magnets and for some species, including red squirrels (Tamiasciurus hudsonicus), American marten (Martes americana), and fisher (Pekania pennanti). Moreover, we were able to capture more species with SiMPL magnets than without, including flying squirrels (Glaucomys spp.), various Cricetidae spp., black bears (Ursus americanus), moose (Alces alces), owls and other birds. The SiMPL wildlife magnet is an effective, low-cost method for surveying wildlife communities, especially rodents and mesocarnivores. It addresses the limited range view presented by other techniques for capturing small mammals on camera traps and enables efficient collection of phenology data, including vegetation and snowpack. This tool has several applications, including monitoring species' responses to silvicultural practices and global change.

Systematic Review Dataset Collection for Intrusion Detection

Open the record for dataset details and reuse information.

Detection of allophane by Zhurong rover indicative of limited water-alteration at Utopia Planitia of Mars

Tianwen-1 mission is China’s first Martian exploration mission, and the Zhurong rover has successfully landed in southern Utopia Planitia. The MarSCoDe onboard the Zhurong rover can obtain mineralogical information on the Martian surface materials by using SWIR spectra. However, the exact types of hydrated phases are not clear in the existing literature. Thus we combined the SWIR spectra and laboratory database of binary mixtures to further constrain mineral types and their contents. Our results reveal that allophane may be the source of hydration features in the SWIR spectra of the Tianwen-1 landing site. This is of great significance to the geological interpretation and process of the landing site.

opencc-by-4.0Mar 2022View details →

Database for "Automatic Detection of Instream Large Wood in Videos Using Deep Learning"

The database contains 21 datasets. Each dataset consists of a folder ('jpgs') with images and a folder  ('txts') with detections according to the YOLO standard.

Navigation Test in Simulated Environment Rosbag. Human obstacle detection.

This repository contains rosbags (ROS 2 Humble) extracted from a navigation test realized in a simulated environment (Amazon Hospital map) with an RB1 robot. The test consist of a navigation form one point to another with a human obstacle avoidance.   This research is part of the project TESCAC, financed by “European Union NextGeneration-EU, the Recovery Plan, Transformation and Resilience, through INCIBE".