Overview

With the rapid advancement of technology, artificial intelligence (AI) and machine learning (ML) have emerged as key drivers of automation, progressively taking over roles traditionally filled by manual inspection methods. Central to this technological revolution is the need for comprehensive datasets, derived from diverse sources like X-ray, SEM (Scanning Electron Microscopy), THz (Terahertz) imaging, acoustic, optical, thermal analysis, and meticulous sample preparation. These datasets are the cornerstone for developing advanced algorithms that are reshaping the future of electronic systems. This scarcity of resources is a significant barrier, hindering innovation and collaborative efforts in the electronic systems sector. Recognizing this crucial need, Physical DB is committed to providing a secure, accessible platform for the sharing and utilization of these essential datasets. Our platform transcends being a mere collection of datasets; it is an active, thriving ecosystem aimed at engaging and empowering professionals across various domains – researchers, engineers, and industry experts alike. We strive to democratize access to these valuable resources, creating a centralized hub for data sharing that nurtures a spirit of collaboration and community.

Optical Imaging

The FPIC dataset was introduced to meet the growing demand for labeled ground truth data in the outsourced PCB manufacturing and hardware assurance sectors read more...

  • • PCB Assurance
  •    
  • • 71,000 annotated instances
  •    
  • • 261 images of 93 separate PCBs

This study examines various methods for automating the recognition and categorization of logos found on printed circuit boards (PCBs) and integrated circuits (ICs). read more...

  • •logos on PCBs and ICs
  •    
  • •18 classes
  •    
  • •>14,000 images

The document titled "Deep Learning-Based Approaches for Text Recognition in PCB Optical Inspection": A Survey read more...

  • •34 high-resolution pcb images used
  •    
  • •6,089 component markings and reference designators
  •    
  • •The annotations are in the ICDAR format.
  •    

The IC SynthLogo dataset is a synthetic image dataset created to enhance the detection of counterfeit and recycled Integrated Circuits (ICs) in Printed Circuit Boards (PCBs). It utilizes read more...

  • •826 number of logos
  •    
  • •188 number of regions to implant
  •    
  • •28436 synthetic images

The document "EC-Seg: Electronic Component Segmentation for PCB Assurance and Counterfeit Avoidance" presents a study focused on enhancing PCB assurance through an automated read more...

  • •Component segmentation and localization
  •    
  • •sensor fusion
  •    
  • •BoM extraction

The DeepIC Logo dataset detailed emerges a crucial response to the growing hardware challenges caused by undetected faults in printed circuit board (PCB) read more...

  • •1010 logo instances
  •    
  • •980 sample images
  •    
  • •119 classes
Datasets

SEM Imaging

The "Explainable Vision System for Hardware Testing and Assurance" (EVHA) is a novel system developed for enhancing the security read more..

      

  • •(CMOS) technology node
  •    
  • •115 images
  •    
  • •3 dwelling times
Datasets

Xray Imaging

The document titled "FICS PCB X-ray: A dataset for automated printed circuit board inter-layers inspection" outlines a significant read more...

  • •5 PCB samples
  •    
  • •16-bit Tiff files
  •    
  • •annotated X-ray layer images
  •    
  • •X-ray data of PCBs

The "PCB X-ray CT Ground Truth Dataset" in the document is a unique and valuable resource for the development and testing of automated read more...

  • •2D X-ray data from 360 degrees
  •    
  • •Converted 16bit Tif to 8-bit JPG
  •    
  • •486 to 1920 Tiff Stack images
  •    
  • • PCB annotations vary (PCB1: 164 vias, 55 traces, 29 pads)

The SegPCBX dataset encompasses 2D reconstructed X-ray images derived from the most optimized X-ray scan, offering extensive research opportunities read more...

  • •COCO JSON format
  •    
  • •400 annotated components
  •    
  • •140 vias, 132 solder balls, 62 pads, 71 traces, and 2 copper planes
  •