Defect Class

Defect classification research focuses on automatically identifying and categorizing flaws in various systems, from manufactured goods to infrastructure components, using advanced computational methods. Current research emphasizes leveraging deep learning architectures, such as Mask R-CNN, RetinaNet, and Vision-Language Models (VLMs) combined with Large Language Models (LLMs), to improve accuracy and robustness, particularly when dealing with limited or imbalanced datasets. These advancements address challenges like insufficient training data and the need for efficient defect detection in diverse modalities (images, videos, sensor data), improving quality control and predictive maintenance across numerous industries. The ultimate goal is to create more reliable and adaptable automated inspection systems that reduce human intervention and improve overall efficiency.

Papers

October 24, 2024

Synth4Seg -- Learning Defect Data Synthesis for Defect Segmentation using Bi-level Optimization
Shancong Mou, Raviteja Vemulapalli, Shiyu Li, Yuxuan Liu, C Thomas, Meng Cao, Haoping Bai, Oncel Tuzel, Ping Huang, Jiulong Shan, Jianjun Shi
Bi Level Optimization Defect Segmentation Defect Class

August 26, 2024

DefectTwin: When LLM Meets Digital Twin for Railway Defect Inspection
Rahatara Ferdousi, M. Anwar Hossain, Chunsheng Yang, Abdulmotaleb El Saddik
Medical LLM Digital Twin Defect Class

April 8, 2024

Progressive Alignment with VLM-LLM Feature to Augment Defect Classification for the ASE Dataset
Chih-Chung Hsu, Chia-Ming Lee, Chun-Hung Sun, Kuang-Ming Wu
Large Vision Language Model Defect Classification Progressive Alignment Defect Sample Defect Class

April 4, 2024

DeepFunction: Deep Metric Learning-based Imbalanced Classification for Diagnosing Threaded Pipe Connection Defects using Functional Data
Yukun Xie, Juan Du, Chen Zhang
Deep Metric Learning Functional Data Analysis Defect Classification Defect Sample Deep Function Pipe Inspection Defect Class

September 12, 2023

Hierarchical Conditional Semi-Paired Image-to-Image Translation For Multi-Task Image Defect Correction On Shopping Websites
Moyan Li, Jinmiao Fu, Shaoyuan Xu, Huidong Liu, Jia Liu, Bryan Wang
Image to Image Translation E Commerce Product Image Unified Image Defect Class Multi Task Image Restoration Image Artifact

November 3, 2022

Deep Learning based Defect classification and detection in SEM images: A Mask R-CNN approach
Bappaditya Dey, Enrique Dehaerne, Kasem Khalil, Sandip Halder, Philippe Leray, Magdy A. Bayoumi
Deep Learning Data Detection SEM Image Defect Classification Mask R CNN Defect Segmentation Semiconductor Defect Inspection Defect Class

July 1, 2022

Few-shot incremental learning in the context of solar cell quality inspection
Julen Balzategui, Luka Eciolaza
Context Information Defect Detection Defect Class Shot Incremental

June 20, 2022

Deep Learning-Based Defect Classification and Detection in SEM Images
Bappaditya Deya, Dipam Goswamif, Sandip Haldera, Kasem Khalilb, Philippe Leraya, Magdy A. Bayoumi
Data Detection SEM Image Defect Classification Defect Class

January 17, 2022

Who supervises the supervisor? Model monitoring in production using deep feature embeddings with applications to workpiece inspection
Michael Banf, Gregor Steinhagen
Feature Embeddings Visual Inspection Production Incident Condition Monitoring Surface Defect Model Monitoring Defect Class Supervisory Performance

Defect Class

Papers

Synth4Seg -- Learning Defect Data Synthesis for Defect Segmentation using Bi-level Optimization

DefectTwin: When LLM Meets Digital Twin for Railway Defect Inspection

Progressive Alignment with VLM-LLM Feature to Augment Defect Classification for the ASE Dataset

DeepFunction: Deep Metric Learning-based Imbalanced Classification for Diagnosing Threaded Pipe Connection Defects using Functional Data

Hierarchical Conditional Semi-Paired Image-to-Image Translation For Multi-Task Image Defect Correction On Shopping Websites

Deep Learning based Defect classification and detection in SEM images: A Mask R-CNN approach

Few-shot incremental learning in the context of solar cell quality inspection

Deep Learning-Based Defect Classification and Detection in SEM Images

Who supervises the supervisor? Model monitoring in production using deep feature embeddings with applications to workpiece inspection