README.md

---
language: en
license: agpl-3.0
library_name: ultralytics
tags:
- military
- drone
- object-detection
- aerial-surveillance
- defense
metrics:
- map
- precision
- recall
model_name: VeritaMilitary
base_model:
- Ultralytics/YOLO26
base_model_relation: finetune
---

# VeritaScan 🛡️
**VeritaScan** is a high-performance, lightweight object detection model specifically optimized for **Aerial Military Surveillance**. It is designed to be deployed on tactical drones with limited onboard computing power.

Developed by **M Mashhudur Rahim (XythicK)** at **Arkito Lab** (A Non-Profit Research Organization).

## 🚀 Model Overview
- **Model Size:** 5.4 MB
- **Target Domain:** Aerial/Drone View
- **Inference Speed:** ~2.7ms per image

## 📊 Training Results
The model was trained to achieve a solid balance between speed and accuracy for edge devices.

| Metric | Value |
| :--- | :--- |
| **mAP@.5-.95** | **88.7%** |
| **Precision** | **96.2%** |
| **Recall** | **98.6%** |

### Performance Visualization
The following graphs illustrate the training progress and model evaluation:

#### Training Curves
![Results](result.jpg) 


## 🎯 Domain Specialization (Aerial vs Ground)
**VeritaMilitary** is highly specialized for top-down perspectives. While it may show sensitivity to perspective shifts in ground-level imagery, it demonstrates exceptional precision in aerial views with confidence scores up to **90%**.

| Aerial Detection (Success) | |
| :---: | :---: |
| ![Aerial](4.png) | ![Ground](5.png) |
| **AirCraft Tracking** |  | **Military Vehicle Tracking** |

## 📦 Multi-Platform Deployment & Downloads
VeritaMilitary supports a wide range of hardware. Click the buttons below to download the optimized formats:

| Framework | Target Hardware | Optimization | One-Click Download |
| :--- | :--- | :--- | :--- |
| <img src="https://files.svgcdn.io/simple-icons/onnx.png" width="40"> | **Universal** (PC, Cloud, Edge) | Standard FP32/FP16 | [**⬇️ Download ONNX**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.onnx) |
| <img src="https://cdn.iconscout.com/icon/free/png-256/free-nvidia-logo-icon-svg-download-png-2945060.png" width="45"> | **NVIDIA Jetson / RTX** | CUDA Accelerated | [**⬇️ Download TensorRT (.engine)**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.engine) |
| <img src="https://1000logos.net/wp-content/uploads/2021/05/Intel-logo.png" width="80"> | **Intel CPU / iGPU** | OpenVINO Runtime | [**⬇️ Download OpenVINO**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan_openvino_model.zip) |
| <img src="https://storage.googleapis.com/gweb-developer-goog-blog-assets/images_archive/original_images/image1_v7xhr8h.png" width="80"> | **Mobile / Android** | Mobile Quantized | [**⬇️ Download TFLite**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan_float16.tflite) |
| <img src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjuUv6hVLci1L1ZfvwIXDkZGetljXlYeUaVyDRyd5SxDf7HtGKrrpQY1Y4vrBWAK9lq9Ezmzw4Vk7v-dNeKyFCg6H_1Lqg43pGbNFHG5spM07j0ThWSPWGzkMdPg0P_KPsFO3-DHfKpR7I/s1600/tensorflowjs.png" width="80"> | **Web Browsers** | TensorFlow.js | [**⬇️ Download TF.js**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan_web_model.zip) |
| <img src="https://www.tensorflow.org/images/tf_logo_horizontal.png" width="80"> | **Legacy Systems** | Frozen Graph (.pb) | [**⬇️ Download TF-PB**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.pb) |
| <img src="https://pytorch.org/assets/images/pytorch-logo.png" width="40"> | **C++ / Production** | TorchScript JIT | [**⬇️ Download TorchScript**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.torchscript) |
| <img src="https://cdn-icons-png.flaticon.com/512/0/747.png" width="35"> | **iOS / macOS** | Apple Neural Engine | [**⬇️ Download CoreML**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.mlpackage.zip) |
| <img src="https://github.com/alibaba/MNN/raw/master/doc/banner.png" width="60"> | **Embedded Devices** | MNN Optimized | [**⬇️ Download MNN**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan.mnn) |
| <img src="https://raw.githubusercontent.com/Tencent/ncnn/master/images/256-ncnn.png" width="40"> | **Embedded Devices** | NCNN Optimized | [**⬇️ Download NCNN**](https://huggingface.co/arkito/VeritaMilitary/resolve/main/veritascan_ncnn_model.zip) |
## 🛠 Usage
```python
from ultralytics import YOLO

# Load VeritaMilitary
model = YOLO('veritascan.pt')

# Run Inference
results = model.predict(source='drone_footage.mp4', imgsz=640, conf=0.25)
results[0].show()
```
### 🏢 About Arkito Lab
Arkito Lab is a non-profit organization dedicated to open-source research and the development of AI solutions for humanitarian and defense technology.

Maintained by: M Mashhudur Rahim (XythicK)
### Cite This Project
If you use this model in your research please cite
```
@software{XythicK_VeritaScan_2026,
  author = {M Mashhudur Rahim (XythicK)},
  title = {VeritaScan: Lightweight Nano Object Detection for Aerial Military Surveillance},
  year = {2026},
  publisher = {Arkito Lab},
  url = {https://huggingface.co/arkito/VeritaScan},
}
```