Datasets:
File size: 2,501 Bytes
908386a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 |
---
license: cc-by-sa-4.0
task_categories:
- image-classification
- image-feature-extraction
- tabular-classification
- feature-extraction
language:
- ar
- en
pretty_name: 'HDL: Hand-drawn Digital Logic Gates Dataset'
---
📌 Dataset Summary
HDL (Hand-drawn Digital Logic Gates) is the first balanced image and tabular dataset dedicated to the classification of hand-drawn digital logic gates. The dataset is designed primarily for machine learning education, computer vision research, and AI-assisted electronics applications.
HDL contains 1,200 hand-drawn images covering the 8 fundamental digital logic gates (AND, OR, NOT (INVERTER), BUFFER, NAND, NOR, XOR, and XNOR), with 150 images per class, ensuring perfect class balance.
🧠 Motivation
Most existing handwritten or hand-drawn datasets focus on:
* Digits (e.g., MNIST)
* Characters and scripts
However, no publicly available balanced dataset existed for hand-drawn digital logic gates, despite their importance in:
* Digital electronics education
* Circuit analysis
* AI-assisted schematic interpretation
HDL was created to fill this gap and to support:
* Teaching Machine Learning to engineering students
* No-code ML experimentation
* Benchmarking shallow and deep learning models
📂 Images Dataset Structure
```
HDL/
├── AND/
├── OR/
├── NOT/
├── BUFFER/
├── NAND/
├── NOR/
├── XOR/
└── XNOR/
```
* Image format: `.jpg` / `.jpeg`
* Image source: Hand-drawn on paper and photographed or scanned
* Contributors: Umm Al-Qura University students + logic gates drawn using Logisim-evolution
📊 Benchmark Results
The dataset has been extensively benchmarked using both no-code tools and deep learning models.
🔹 Best Result
* Model: EfficientNet with data augmentation
* Platform: Liner.ai
* Accuracy: 92.5%
* F1-score: 92.5%
🔹 Feature Extraction (Image Embedding) (Tabular Dataset)
Image embeddings were generated using:
* SqueezeNet
* Inception v3
* Google ViT
* Facebook DINOv2
* Nvidia MambaVision
Top-performing classifiers:
* Logistic Regression
* Neural Networks
🧪 Typical Use Cases
HDL is suitable for:
🎓 Machine Learning Education
🧠 TinyML & Edge AI
✍️ OCR-like systems for circuit digitization
🧩 Logic gate classification
🖼️ Image embedding & feature extraction experiments
🛠️ No-code ML platforms (Orange Data Mining, Liner.ai)
📄 Research Paper: To be published soon.
Please cite the dataset if used in academic work. |