# DigitalSharp — "Jinjing" Professional Blurred Image & Video Restoration System V3.1.0

> "Make every blurred frame reveal its truth" — a professional-grade image and video enhancement solution for public security, judicial, criminal investigation and surveillance.

---

## 1. Product Positioning

DigitalSharp integrates neural-network AI models with classical digital image-processing algorithms to provide a professional image/video restoration system tailored for forensic agencies, police investigations, prosecutors, surveillance operations and archival digitization.

Addressing common forensic challenges such as blur from motion, shake, defocus, low light, heavy noise and low resolution, DigitalSharp offers workflows from one-click automatic enhancement to expert parameter tuning and produces court-ready processing reports that ensure traceability and reproducibility.

---

## 2. Core Advantages

1. Forensic-grade positioning (not consumer beautification)
- Generates traceable processing reports (source, result, algorithm chain, parameters, hashes)
- Records every parameter and algorithm call to meet evidence-chain compliance
- Supports ROI-focused enhancement while preserving original evidence integrity

2. Dual engine: AI + classical algorithms
- Deep-learning models developed over years for intelligent inference
- Classical algorithms: Wiener, Richardson–Lucy, blind deconvolution, Shock filter, Tikhonov-TV, Fourier domain enhancement, Retinex
- Both engines can be chained as processing pipelines to handle complex degradations

3. GPU acceleration
- Supports GPU acceleration across platforms (NVIDIA / AMD / Intel)
- Minimal configuration; out-of-the-box use
- Video batch processing can be 5–20× faster than CPU

4. AutoEnhance intelligent one-click
- Automatic blur-type detection (motion/defocus/noise/low‑light)
- Auto-selects optimal algorithm chain for novices to achieve expert-level results
- Scene classifier to detect text/face/license‑plate/texture

5. End-to-end video capabilities
- Frame-by-frame AI deblur / super-resolution
- Multi-frame averaging to improve SNR and reveal the clearest moment
- Keyframe extraction, GIF export, batch queue processing

6. Professional toolbox
- Compare slider for before/after
- Pixel‑level zoom/magnifier
- Image-quality metrics (PSNR / SSIM / Laplacian variance)
- Full history stack, undo/redo, and preset management

---

## 3. Core Features Matrix

AI enhancement modules
- Blind deblurring: neural models for motion/shake/defocus
- Defocus repair: models for lens defocus
- Super-resolution: SR models for low-resolution footage
- Intelligent denoising: dedicated denoising chains for night scenes
- Face enhancement: face restoration models
- Text enhancement: models for license plates, documents
- Texture enhancement: fingerprint, trace, hair texture recovery

Classical toolbox
- Sharpening: USM, Laplacian, guided sharpen
- Deconvolution: Wiener, Richardson–Lucy, blind deconv, Tikhonov-TV
- Frequency tools: Fourier visualization, bandpass/notch filters
- Color/exposure: Levels, Curves, White Balance, HSL, channel ops
- Geometry: perspective correction, crop, ROI selection
- Morphology: dilation/erosion/open/close, edge detection
- Dehaze/enhance: Retinex, dark-channel dehaze, Shock filter

Batch & engineering
- Batch processing dialog and chainable pipelines
- Video frame processing for MP4/AVI/MOV
- GIF export for comparison demos
- Recent files and drag-and-drop support

Forensic report generation
- Produce PDF/HTML evidence reports with source image, processed result, complete parameter chain, timestamps, operator and file hashes

---

## 4. Typical Use Cases

- Police investigations: suspect face restoration, license‑plate enhancement, low-light denoising
- Forensic examination: authentication and enhancement for evidence materials
- Surveillance operations: CCTV analysis, traffic accident reconstruction
- Archival restoration: historical photo and archive digitization

---

## 5. Technical Highlights

| Dimension | DigitalSharp | Consumer image editor |
|---|---|---|
| Algorithm types | AI + classical dual engine | Single filters |
| GPU acceleration | Cross-platform DirectML | CUDA-only or none |
| Forensic compliance | Built-in report generation | None |
| Pipelines | Reproducible, saveable pipelines | Single-step edits |
| Video support | Frame-by-frame AI + multi-frame fusion | Weak or absent |
| Model extensibility | Hot model loading | Closed |
| Professional tools | PSNR/SSIM/spectrum/ROI | Few or none |

---

## 6. Product Specs

- Platform: Windows 10 / 11 x64
- Recommended:
- CPU: Intel i5 / AMD Ryzen 5 or better
- Memory: 16 GB+
- GPU: DirectX 12 compatible discrete GPU (4 GB VRAM recommended)
- Storage: 10 GB free (including model library)
- Stack: Qt 6 + OpenCV + ONNX Runtime (DirectML)
- Licensing: machine-bound (one machine key), 30-day trial available

---

## 7. Customer Value

- Efficiency: one-click processing compresses hours of manual tuning into minutes
- Compliance: reports suitable as attachments to forensic opinions
- Reliability: combined deep learning and classical algorithms
- Security: fully local processing — no cloud upload
- Sustainability: plugin model ecosystem for continuous improvement

---

## 8. Contact

- Business: business@digitalsharp.cn
- Website: https://www.iDigitalForensics.cn
- Support: 182-0599-8855

---

DigitalSharp — providing a clearer eye for the truth.

*© iDigitalForensics. All Rights Reserved.*