磁驱动软胶囊内镜的定位、校准、控制与设计

数据集获取（Figshare）

• 源站标题：Localization, Calibration, Control, and Design of Magnetically Actuated Soft Capsule Endoscopes
• DOI：10.1184/R1/7584005.v1
• 发布日期：2019-01-18
• 文件数：5 个
• 文件总体积：25.55 MB
• 源站页面：https://kilthub.cmu.edu/articles/dataset/Localization_Calibration_Control_and_Design_of_Magnetically_Actuated_Soft_Capsule_Endoscopes/7584005

命令行下载

curl -L -o "Son_cmu_0041E_10344.pdf" "https://ndownloader.figshare.com/files/14083637"
curl -L -o "S1.mp4" "https://ndownloader.figshare.com/files/14083727"
curl -L -o "s2.mp4" "https://ndownloader.figshare.com/files/14083730"
curl -L -o "S3.mp4" "https://ndownloader.figshare.com/files/14083733"
curl -L -o "S4.mp4" "https://ndownloader.figshare.com/files/14083736"

文件清单

• Son_cmu_0041E_10344.pdf（3.99 MB）
• S1.mp4（1.86 MB）
• s2.mp4（2.99 MB）
• S3.mp4（7.72 MB）
• S4.mp4（8.98 MB）

文件总体积 25.55 MB（源站 API 实测）。

数据加载示例（图像类）

from PIL import Image
import glob, os

files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)

# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path)  # 要求 子目录=类别

目录组织与标注格式以源站说明和下载后实际文件为准。

数据说明

Current medical technologies are converging to minimally invasive diagnosis and therapy.The effort to reduce patient discomfort in gastrointestinal (GI) tract diagnoses resulted in thedevelopment of wireless capsule endoscopes (WCEs). In the form of a pill, a camera carryingWCE travels through the GI tract by natural peristalsis while it collects images of the internalwall of the GI tract. The operation might be entirely painless. However, their inability to haveactive locomotion and control limit detailed diagnoses, therapeutic functions, and minimizationof the operation time. Actively controlled WCEs would resolve those challenges.This thesis provides methods for the active control of WCEs using magnetic interactions,and applies those methods to a robotic biopsy capsule endoscope. First, a localization methodfor meso-scale magnetic robots is developed. The method utilizes a magnetic sensor arraywhere a magnetically actuated capsule endoscope (MACE) does not require a special devicefor localization but a single magnet. The method is beneficial to reduce the size and the batteryconsumption of the MACE. The method focuses on decoupling the magnetic field of theMACE from the magnetic field of the actuator, and developing a real-time localization algorithm.Second, an automatic calibration method for magnetic actuation and sensing systemsis presented. The method calibrates a number of nonlinear magnetic sensors and a number ofelectromagnets. The method is capable of calibrating 1.8k parameters in an exemplary systemin a reasonable time without human labor. In this work, Bundle Adjustment framework fromComputer Vision is modified and adapted to magnetic robot sensing and actuation systems.This work would be useful for a magnetic system which requires frequent reconfiguration orsensor/actuator gain updates. Third, control methods for a meso-scale magnetic robots on asurface with non-uniform magnetic field actuations are presented. The control methods utilizemagnetic energy wells to cope with a low actuation bandwidth compared to the fast dynamicsof the capsule endoscope. Additionally, we present a teleoperation system to mitigate the orientationcoordination difficulty

许可

In Copyright

引用

学术使用请引用 DOI 10.1184/R1/7584005.v1（source: https://kilthub.cmu.edu/articles/dataset/Localization_Calibration_Control_and_Design_of_Magnetically_Actuated_Soft_Capsule_Endoscopes/7584005）。