CLICK ON THE NAME OF THE DATASET TO OPEN THE RELEVANT LINK & DOWNLOAD
 |
"TARRAGONA" GRAPH MATCHING: This repository contains 7 datasets where the structure
consists of registers containing a pair of graphs, their classes, and a
labelling/matching/correspondence between them. PLEASE REFERENCE AS: C.F. Moreno-Garcia, X. Cortés & F. Serratosa, A
Graph Repository for Learning Error-Tolerant Graph Matching, Structural,
Syntactic, and Statistical Pattern Recognition, Vol 10029, 2016, pp.519-529. https://doi.org/10.1007/11815921.
|
 |
P&ID SYMBOL DATASET 1: This dataset contains 1'187 symbols from engineering
drawings split into 37 classes. Each symbol is represented as a 100x100 pixel
matrix (flattened as a 1x10'000 vector in the .csv file) with the class in
the final column. PLEASE REFERENCE AS: E. Elyan, C.F. Moreno-Garcia & C. Jayne, Symbols
classification in engineering drawings, International Joint Conference on
Neural Networks, 2018, https://doi.org/10.1109/IJCNN.2018.8489087.
|
 |
SYMBOLS IN ENGINEERING DRAWINGS (SiED): This dataset contains 2'432 symbols from
engineering drawings split into 39 classes. Each symbol is represented as a
100x100 pixel matrix (flattened as a 1x10'000 vector in the .csv file) with
the class in the final column. PLEASE REFERENCE AS: E. Elyan, C.F. Moreno-García & P. Johnston,
Symbols in Engineering Drawings (SiED): An Imbalanced Dataset Benchmarked by
Convolutional Neural Networks, Engineering Applications of Neural Networks,
2020: pp. 215-224, https://doi.org/10.1007/978-3-030-48791-1.
|
 |
Corrosion segmentation in underwater images: This dataset is composed of two version. The first one (called all_images_no_masks) has 24 underwater images with corrosion, and 1233 surface images (128 with no corrosion and 1105 with corrosion). There second one (called train_with_masks) has 1068 images of surface and underwater images combined, but this time with the mask segmentations contained in a file called annotations.json. PLEASE REFERENCE AS: C. Pirie & C. F. Moreno-García, Image Pre-processing and Segmentation for Real-Time Subsea Corrosion Inspection. In: Engineering Applications of Neural Networks. 2021. p. 220–31. https://link.springer.com/10.1007/978-3-030-80568-5_19 |
 |
CARDIAC images:
This dataset is a subset of 100 images extracted from: EchoNet-Dynamic Cardiac Ultrasound | Center for Artificial Intelligence in Medicine Imaging;. Available from: https://aimi.stanford.edu/echonet-dynamic-cardiac-ultrasound. PLEASE REFERENCE AS: A. Cervantes-Guzmán, K. McPherson, J. Olveres, C. F. Moreno-García, F. T. Robles, E. Elyan, B. Escalante-Ramírez, “Robust cardiac segmentation corrected with heuristics”, PLOS ONE 18(10): e0293560. https://doi.org/10.1371/journal.pone.0293560.
|
More datasets available at my GitHub and Kaggle sites. Moreover, more code and datasets produced by our research groups can be found here.
