---
license: apache-2.0
base_model:
- microsoft/conditional-detr-resnet-50
pipeline_tag: object-detection
datasets:
- tech4humans/signature-detection
metrics:
- f1
- precision
- recall
library_name: transformers
inference: false
tags:
- object-detection
- signature-detection
- detr
- conditional-detr
- pytorch
model-index:
- name: tech4humans/conditional-detr-50-signature-detector
results:
- task:
type: object-detection
dataset:
type: tech4humans/signature-detection
name: tech4humans/signature-detection
split: test
metrics:
- type: precision
value: 0.936524
name: mAP@0.5
- type: precision
value: 0.653321
name: mAP@0.5:0.95
---
# **Conditional-DETR ResNet-50 - Handwritten Signature Detection**
This repository presents a Conditional-DETR model with ResNet-50 backbone, fine-tuned to detect handwritten signatures in document images. This model achieved the **highest mAP@0.5 (93.65%)** among all tested architectures in our comprehensive evaluation.
| Resource | Links / Badges | Details |
|---------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **Article** | [](https://huggingface.co/blog/samuellimabraz/signature-detection-model) | A detailed community article covering the full development process of the project |
| **Model Files (YOLOv8s)** | [](https://huggingface.co/tech4humans/yolov8s-signature-detector) | **Available formats:** [](https://pytorch.org/) [](https://onnx.ai/) [](https://developer.nvidia.com/tensorrt) |
| **Dataset – Original** | [](https://universe.roboflow.com/tech-ysdkk/signature-detection-hlx8j) | 2,819 document images annotated with signature coordinates |
| **Dataset – Processed** | [](https://huggingface.co/datasets/tech4humans/signature-detection) | Augmented and pre-processed version (640px) for model training |
| **Notebooks – Model Experiments** | [](https://colab.research.google.com/drive/1wSySw_zwyuv6XSaGmkngI4dwbj-hR4ix) [](https://api.wandb.ai/links/samuel-lima-tech4humans/30cmrkp8) | Complete training and evaluation pipeline with selection among different architectures (yolo, detr, rt-detr, conditional-detr, yolos) |
| **Notebooks – HP Tuning** | [](https://colab.research.google.com/drive/1wSySw_zwyuv6XSaGmkngI4dwbj-hR4ix) [](https://api.wandb.ai/links/samuel-lima-tech4humans/31a6zhb1) | Optuna trials for optimizing the precision/recall balance |
| **Inference Server** | [](https://github.com/tech4ai/t4ai-signature-detect-server) | Complete deployment and inference pipeline with Triton Inference Server
[](https://docs.openvino.ai/2025/index.html) [](https://www.docker.com/) [](https://developer.nvidia.com/triton-inference-server) |
| **Live Demo** | [](https://huggingface.co/spaces/tech4humans/signature-detection) | Graphical interface with real-time inference
[](https://www.gradio.app/) [](https://plotly.com/python/) |
---
## **Dataset**
The training utilized a dataset built from two public datasets: [Tobacco800](https://paperswithcode.com/dataset/tobacco-800) and [signatures-xc8up](https://universe.roboflow.com/roboflow-100/signatures-xc8up), unified and processed in [Roboflow](https://roboflow.com/).
**Dataset Summary:**
- Training: 1,980 images (70%)
- Validation: 420 images (15%)
- Testing: 419 images (15%)
- Format: COCO JSON
- Resolution: 640x640 pixels
---
## **Training Process**
The training process involved the following steps:
### 1. **Model Selection:**
Various object detection models were evaluated to identify the best balance between precision, recall, and inference time.
| **Metric** | [rtdetr-l](https://github.com/ultralytics/assets/releases/download/v8.2.0/rtdetr-l.pt) | [yolos-base](https://huggingface.co/hustvl/yolos-base) | [yolos-tiny](https://huggingface.co/hustvl/yolos-tiny) | [conditional-detr-resnet-50](https://huggingface.co/microsoft/conditional-detr-resnet-50) | [detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) | [yolov8x](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8x.pt) | [yolov8l](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8l.pt) | [yolov8m](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8m.pt) | [yolov8s](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8s.pt) | [yolov8n](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8n.pt) | [yolo11x](https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11x.pt) | [yolo11l](https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11l.pt) | [yolo11m](https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11m.pt) | [yolo11s](https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11s.pt) | [yolo11n](https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt) | [yolov10x](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10x.pt) | [yolov10l](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10l.pt) | [yolov10b](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10b.pt) | [yolov10m](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10m.pt) | [yolov10s](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10s.pt) | [yolov10n](https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10n.pt) |
|:---------------------|---------:|-----------:|-----------:|---------------------------:|---------------:|--------:|--------:|--------:|--------:|--------:|--------:|--------:|--------:|--------:|--------:|---------:|---------:|---------:|---------:|---------:|---------:|
| **Inference Time - CPU (ms)** | 583.608 | 1706.49 | 265.346 | 476.831 | 425.649 | 1259.47 | 871.329 | 401.183 | 216.6 | 110.442 | 1016.68 | 518.147 | 381.652 | 179.792 | 106.656 | 821.183 | 580.767 | 473.109 | 320.12 | 150.076 | **73.8596** |
| **mAP50** | 0.92709 | 0.901154 | 0.869814 | **0.936524** | 0.88885 | 0.794237| 0.800312| 0.875322| 0.874721| 0.816089| 0.667074| 0.707409| 0.809557| 0.835605| 0.813799| 0.681023| 0.726802| 0.789835| 0.787688| 0.663877| 0.734332 |
| **mAP50-95** | 0.622364 | 0.583569 | 0.469064 | 0.653321 | 0.579428 | 0.552919| 0.593976| **0.665495**| 0.65457 | 0.623963| 0.482289| 0.499126| 0.600797| 0.638849| 0.617496| 0.474535| 0.522654| 0.578874| 0.581259| 0.473857| 0.552704 |
#### Highlights:
- **Best mAP50:** `conditional-detr-resnet-50` (**0.936524**)
- **Best mAP50-95:** `yolov8m` (**0.665495**)
- **Fastest Inference Time:** `yolov10n` (**73.8596 ms**)
Detailed experiments are available on [**Weights & Biases**](https://api.wandb.ai/links/samuel-lima-tech4humans/30cmrkp8).
### 2. **Hyperparameter Tuning:**
The YOLOv8s model, which demonstrated a good balance of inference time, precision, and recall, was selected for hyperparameter tuning.
[Optuna](https://optuna.org/) was used for 20 optimization trials.
The hyperparameter tuning used the following parameter configuration:
```python
dropout = trial.suggest_float("dropout", 0.0, 0.5, step=0.1)
lr0 = trial.suggest_float("lr0", 1e-5, 1e-1, log=True)
box = trial.suggest_float("box", 3.0, 7.0, step=1.0)
cls = trial.suggest_float("cls", 0.5, 1.5, step=0.2)
opt = trial.suggest_categorical("optimizer", ["AdamW", "RMSProp"])
```
Results can be visualized here: [**Hypertuning Experiment**](https://api.wandb.ai/links/samuel-lima-tech4humans/31a6zhb1).
### 3. **Evaluation:**
The models were evaluated on the test set at the end of training in ONNX (CPU) and TensorRT (GPU - T4) formats. Performance metrics included precision, recall, mAP50, and mAP50-95.
#### Results Comparison:
| Metric | Base Model | Best Trial (#10) | Difference |
|------------|------------|-------------------|-------------|
| mAP50 | 87.47% | **95.75%** | +8.28% |
| mAP50-95 | 65.46% | **66.26%** | +0.81% |
| Precision | **97.23%** | 95.61% | -1.63% |
| Recall | 76.16% | **91.21%** | +15.05% |
| F1-score | 85.42% | **93.36%** | +7.94% |
---
## **Results**
After hyperparameter tuning of the YOLOv8s model, the best model achieved the following results on the test set:
- **Precision:** 94.74%
- **Recall:** 89.72%
- **mAP@50:** 94.50%
- **mAP@50-95:** 67.35%
- **Inference Time:**
- **ONNX Runtime (CPU):** 171.56 ms
- **TensorRT (GPU - T4):** 7.657 ms
---
## **How to Use**
### **Installation**
```bash
pip install transformers torch torchvision pillow
```
### **Inference**
```python
from transformers import AutoImageProcessor, AutoModelForObjectDetection
from PIL import Image
import torch
# Load model and processor
model_name = "tech4humans/conditional-detr-50-signature-detector"
processor = AutoImageProcessor.from_pretrained(model_name)
model = AutoModelForObjectDetection.from_pretrained(model_name)
# Load and process image
image = Image.open("path/to/your/document.jpg")
inputs = processor(images=image, return_tensors="pt")
# Run inference
with torch.no_grad():
outputs = model(**inputs)
# Post-process results
target_sizes = torch.tensor([image.size[::-1]])
results = processor.post_process_object_detection(
outputs, target_sizes=target_sizes, threshold=0.5
)[0]
# Extract detections
for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
box = [round(i, 2) for i in box.tolist()]
print(f"Detected signature with confidence {round(score.item(), 3)} at location {box}")
```
### **Visualization**
```python
import matplotlib.pyplot as plt
import matplotlib.patches as patches
from PIL import Image
def visualize_predictions(image_path, results, threshold=0.5):
image = Image.open(image_path)
fig, ax = plt.subplots(1, figsize=(12, 9))
ax.imshow(image)
for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
if score > threshold:
x, y, x2, y2 = box.tolist()
width, height = x2 - x, y2 - y
rect = patches.Rectangle(
(x, y), width, height,
linewidth=2, edgecolor='red', facecolor='none'
)
ax.add_patch(rect)
ax.text(x, y-10, f'Signature: {score:.3f}',
bbox=dict(boxstyle="round,pad=0.3", facecolor="yellow", alpha=0.7))
ax.set_title("Signature Detection Results")
plt.axis('off')
plt.show()
# Use the visualization
visualize_predictions("path/to/your/document.jpg", results)
```
---
## **Demo**
You can explore the model and test real-time inference in the Hugging Face Spaces demo, built with Gradio and ONNXRuntime.
[](https://huggingface.co/spaces/tech4humans/signature-detection)
---
## 🔗 **Inference with Triton Server**
If you want to deploy this signature detection model in a production environment, check out our inference server repository based on the NVIDIA Triton Inference Server.
---
## **Infrastructure**
### Software
The model was trained and tuned using a Jupyter Notebook environment.
- **Operating System:** Ubuntu 22.04
- **Python:** 3.10.12
- **PyTorch:** 2.5.1+cu121
- **Ultralytics:** 8.3.58
- **Roboflow:** 1.1.50
- **Optuna:** 4.1.0
- **ONNX Runtime:** 1.20.1
- **TensorRT:** 10.7.0
### Hardware
Training was performed on a Google Cloud Platform n1-standard-8 instance with the following specifications:
- **CPU:** 8 vCPUs
- **GPU:** NVIDIA Tesla T4
---
## **License**
### Model Weights, Code and Training Materials – **Apache 2.0**
- **License:** Apache License 2.0
- **Usage:** All training scripts, deployment code, and usage instructions are licensed under the Apache 2.0 license.
---
## **Contact and Information**
For further information, questions, or contributions, contact us at **iag@tech4h.com.br**.
## **Author**
Samuel Lima
AI Research Engineer
|
Responsibilities in this Project
- 🔬 Model development and training
- 📊 Dataset analysis and processing
- ⚙️ Hyperparameter optimization and performance evaluation
- 📝 Technical documentation and model card
|
---
