Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,35 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- computer-vision
|
| 5 |
+
- circuit-recognition
|
| 6 |
+
- deep-learning
|
| 7 |
+
- object-detection
|
| 8 |
+
- pose-estimation
|
| 9 |
+
pipeline_tag: image-classification
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# HCD Circuit Recognition Weights
|
| 13 |
+
|
| 14 |
+
This repository hosts the pretrained weights for the paper:
|
| 15 |
+
|
| 16 |
+
**Deep Learning-Based Recognition and Structural Parsing of Hand-Drawn Circuit Diagrams**
|
| 17 |
+
|
| 18 |
+
## Components
|
| 19 |
+
|
| 20 |
+
The repository contains pretrained weights for the following models used in the pipeline:
|
| 21 |
+
|
| 22 |
+
- YOLOv10 component detector
|
| 23 |
+
- PARSeq text recognition model
|
| 24 |
+
- HAWP wire/junction detection model
|
| 25 |
+
- DINOv2 component subtype classification model
|
| 26 |
+
- ViTPose endpoint semantic inference model
|
| 27 |
+
|
| 28 |
+
## Files
|
| 29 |
+
|
| 30 |
+
- `weights_yolo_parseq.zip` – YOLOv10 + PARSeq weights
|
| 31 |
+
- `weights_hawp.zip` – HAWP junction detection weights
|
| 32 |
+
- `weights_dinov2.zip` – DINOv2 component subtype classifier weights
|
| 33 |
+
- `weights_vitpose.zip` – ViTPose endpoint semantic inference weights
|
| 34 |
+
|
| 35 |
+
SHA256 checksums are provided in `SHA256SUMS.txt`.
|