we introduce our Hard Defect Classification Dataset (HDCD), which is primarily constructed based on the Multi Classifier Data (MCD) [38], a multi-source dataset comprising self-collected image samples along with contributions from several Departments of transportation inspection databases. Initially, we integrated data from diverse sources based on the defect classification criteria of the MCD. Subsequently, we partitioned the dataset into a training set and a test set. To simulate real-world challenges, we specifically curated the test set to include hard samples

 

Dataset Files

You must be an IEEE Dataport Subscriber to access these files. Subscribe now or login.

[1] Shiyu Tao, "HDCD", IEEE Dataport, 2025. [Online]. Available: http://dx.doi.org/10.21227/dpsp-4w08. Accessed: Feb. 17, 2025.
@data{dpsp-4w08-25,
doi = {10.21227/dpsp-4w08},
url = {http://dx.doi.org/10.21227/dpsp-4w08},
author = {Shiyu Tao },
publisher = {IEEE Dataport},
title = {HDCD},
year = {2025} }
TY - DATA
T1 - HDCD
AU - Shiyu Tao
PY - 2025
PB - IEEE Dataport
UR - 10.21227/dpsp-4w08
ER -
Shiyu Tao. (2025). HDCD. IEEE Dataport. http://dx.doi.org/10.21227/dpsp-4w08
Shiyu Tao, 2025. HDCD. Available at: http://dx.doi.org/10.21227/dpsp-4w08.
Shiyu Tao. (2025). "HDCD." Web.
1. Shiyu Tao. HDCD [Internet]. IEEE Dataport; 2025. Available from : http://dx.doi.org/10.21227/dpsp-4w08
Shiyu Tao. "HDCD." doi: 10.21227/dpsp-4w08