SecureMark

SecureMark is a professional Flutter application designed to secure documents and images with visible watermarks and advanced invisible steganography before sharing them.

The project is designed for high-security document distribution: choose files, apply multi-layered protection, preview results, and securely share them.

Core Features Overview

• Settings Profiles: One-tap presets (Custom, Identity, Image, Doc, QR Code) for instant workflow configuration.
• AI Protection Suite:
  • Anti-AI Removal Protection: Position and alpha jitter to prevent AI from learning and erasing watermarks.
  • AI Cloaking (Adversarial): Injects invisible noise to disrupt AI style extraction and block OCR scrapers.
• Multi-layer Steganography: 3 independent invisible methods (LSB text, hidden file, and robust DCT).
• Advanced File Management: Manage large batches with a dedicated selection modal featuring PDF first-page previews and easy removal.
• Batch Processing: Background threading (Isolates) with linear, granular progress tracking.
• Built-in Analyzer: Detect and extract hidden signatures or embedded files with AES-256 decryption.
• Cross-platform: Android share intents, iOS export, and desktop drag-and-drop support.

User Interface & Workflow

Main Interface

• Profile Selector: High-visibility horizontal chips for rapid switching between security presets.
• Interactive Preview:
  • Smart Zoom: Double-tap to zoom at specific coordinates for detailed inspection.
  • A/B Toggle: Quickly compare the original vs. protected result (images only).
  • Live Shader: Real-time preview of transparency and color effects.
• Merged Controls: Status icons are now integrated directly into the styling cards to maximize vertical space.

File Input & Management

• Selected Files Modal: Click the "Selected files" label to open a grid view of all pending files with thumbnails and individual "remove" buttons.
• Fast PDF Previews: Thumbnails for PDF files are generated instantly using the first page.
• Loading Feedback: Immediate visual feedback during file selection and preparation.
• Input Variety: Support for File Picker, Scan Document (with edge detection), Desktop Drag & Drop, and Android Share Intents.

Privacy & Offline Document Scanning

SecureMark prioritizes user privacy by processing all documents locally on the device.

• Offline Processing: Document scanning, edge detection, and perspective correction are performed entirely on-device using native system APIs (Google Play Services Document Scanner on Android and VisionKit on iOS).
• "Secure by Google" Branding: On Android, the scanner uses official Google Play Services APIs. The "Secure by Google" badge indicates that the camera and edge-detection code is a verified, official system component.
• On-Device ML: All machine learning models for document detection are downloaded to your phone. Document data is never uploaded to the cloud or sent to external servers, ensuring your sensitive files remain strictly private.

Specialized Dialogs

1. File Analyzer: Detect and extract hidden signatures, robust watermarks, and embedded files
2. Steganography Options: Configure LSB text signature and hidden file embedding with AES-256 encryption
3. QR Watermark Options: Configure QR code type, position, size, and opacity
4. Expert Options: Advanced settings for quality control, anti-AI protection, and output customization
5. Logs Viewer: Real-time application logs (100 most recent) with export capability
6. About Dialog: Version information and project details
7. Processing Progress: Real-time progress updates during batch operations
8. Save Results: File naming and location selection with preview

Watermarking Capabilities

Visible Watermarks

Before:

After - Text Watermark:

• Text Watermarks: Custom text with automatic timestamp injection
• Image/Logo Watermarks: Brand logo overlay with preserved aspect ratio

• Transparency Control: 0-100% opacity adjustment
• Density Control: Adaptive placement algorithm (8-max instances based on image dimensions)
• Rotation: Random 15° step rotations for anti-pattern AI protection
• Color Options:
  • Random HSV color generation with 6 color variations per watermark
  • Custom color picker with RGB selection
  • Per-watermark color variation support

Smart Placement Algorithm

• Grid-based Positioning: Automatic cell calculation (rows × columns) based on image size
• Random Jitter: Positioning variation within cells for natural appearance
• Collision Avoidance: Retry logic (6 attempts per cell, 12 global) to prevent overlap
• Adaptive Density: Watermark count scales with image dimensions

Font Rendering

• Bitmap Font Support: Arial rendered at 14pt/24pt/48pt for system compatibility
• TrueType Rendering: Flutter canvas-based rendering for all other fonts
• Stamp Caching: Pre-rendered watermark stamps cached per font-size combination
• Colorization: Efficient colorization at composite time

PDF Support

• Vector Watermarking: Native PDF graphics layer insertion (background mode) preserves text quality
• Raster Fallback: Automatic page-by-page rasterization for malformed PDFs
• Multi-page Processing: Iterate through all pages in PDF documents
• Preview Generation: First page rendered at 72 DPI for preview display

Advanced Security & Steganography

SecureMark implements three independent invisible protection layers.

1. LSB Text Signature

Purpose: Embed invisible text signature for ownership verification

2. Hidden File Embedding

Purpose: Embed entire secondary files (documents, images, archives) within the image

3. Robust DCT Watermarking

Purpose: Persistent signature that survives image transformations (compression, resizing, etc.)

Unified File Analyzer

The File Analyzer scans to find Steganography.

QR Code Integration

SecureMark can embed QR codes watermarks with three content types:

QR Code Types

1. Metadata (JSON)

Standard SecureMark metadata format:

• Author: Document creator name
• URL: Reference website or source
• Timestamp: ISO 8601 format generation time
• App Info: SecureMark identifier and version
• Toggles: Individual field enable/disable
• Format: Compact JSON encoding

2. URL (Website Redirect)

Direct website link encoding:

• Use Case: Product pages, documentation, verification portals
• Format: Plain URL string (https://...)
• Scanning: Opens directly in mobile browsers

3. vCard (Contact Information)

vCard 3.0 format for professional contact sharing:

• Fields:
  • Full Name (First + Last)
  • Organization
  • Phone Number (TYPE=CELL)
  • Email Address (TYPE=INTERNET)
  • Website URL
  • Physical Address
• Format: Standard vCard 3.0 specification
• Compatibility: Works with all major contact apps (iOS Contacts, Google Contacts, Outlook)

Visual Configuration

• Position: 5 options (Top-Left, Top-Right, Bottom-Left, Bottom-Right, Center)
• Size: 50-200 pixels (scalable)
• Opacity: 0.0-1.0 (fully transparent to fully opaque)
• Margin: 20-pixel automatic margin from edges
• Error Correction: Level H (High - 30% recovery)
• Mode Toggle: Visible QR watermark or invisible (metadata-only)

Font System

SecureMark provides 22 font options from three sources with intelligent loading and caching:

System Fonts (1 option)

• Arial (System Default) - Bitmap rendering at 3 sizes (14pt, 24pt, 48pt)

Google Fonts (9 options)

Dynamically downloaded from Google Fonts API:

• Roboto (Modern, sans-serif)
• Open Sans (Clean, highly legible)
• Lato (Professional, business-friendly)
• Montserrat (Bold, geometric)
• Poppins (Rounded, friendly)
• Noto Sans (Universal, multi-language support)
• Source Code Pro (Monospace, technical)
• Playfair Display (Elegant, serif)
• Oswald (Strong, condensed)

Custom TTF Fonts (6 options)

Pre-bundled in application assets:

• Charis SIL (Serif, academic)
• Liberation Mono (Monospace, open-source)
• Liberation Serif (Traditional, book-style)
• Bitstream Vera Sans (Classic, clean)
• Custom Roboto (Local TTF version)
• Custom Open Sans (Local TTF version)

Font Rendering Architecture

• Bitmap Rendering: Arial only, using image package's built-in bitmap fonts (14/24/48pt)
• TrueType Rendering: All other fonts rendered via Flutter Canvas API
  • Pre-rendered stamps cached per font-size combination
  • Colorized at composite time for efficiency
  • Fallback to bitmap if TTF rendering fails

Font Caching System

• LRU Cache: Least Recently Used eviction strategy
• Maximum Cache Size: 50 MB
• Cache Scope: Font bytes stored per font (not per size)
• Pre-loading: Common fonts (Roboto, Open Sans) pre-loaded at startup
• Sources:
  1. System bitmap (immediate)
  2. Local assets (fast, bundled TTF files)
  3. Google Fonts API (network download, cached after first use)

How It Works

1. Enter the watermark text or configure a QR code.
2. Pick files (images or PDFs) or drag and drop them.
3. (Optional) Configure Steganography to hide secret data or files.
4. Adjust visibility settings (Transparency, Density, Font).
5. Click Apply SecureMark.
6. Preview the result (with A/B comparison) and verify the hidden signatures.
7. Save or share the protected outputs.

Export & Processing Options

Batch Processing

• Background Isolation: All heavy processing runs in Dart Isolates (separate threads) using SendPort for safe communication.
• Linear Progress Tracking: Completely redesigned progress bar showing granular steps (Loading, Anti-AI, Watermarking, etc.).
• Multilingual Support: All progress status messages are fully localized (EN, FR, DE, IT).
• Cancellation Token: Graceful cancellation support during processing.
• Result Caching: LRU cache (max 10 results) to avoid reprocessing identical requests.

Quality Control

• JPEG Quality: Adjustable 1-100 (default: 75)
• WebP Quality: Adjustable 1-100 (default: 75)
• PNG Compression: Level 2 (balanced speed/size) for lossless steganography
• Format Preservation:
  • HEIC/HEIF → JPEG conversion
  • Original format maintained when possible
  • Forced PNG output when steganography is enabled (lossless requirement)

Target Size & Resizing

• Auto-resize: Intelligent downscaling for mobile-friendly sharing
• Target Size: User-defined maximum dimension (e.g., 2048px)
• Aspect Ratio: Always preserved during resize
• Interpolation: Average interpolation for quality preservation
• Memory Protection: Automatic fallback on memory limit exceeded errors

Output Customization

• File Prefix: Customizable (default: securemark-)
• Timestamp Injection:
  • Optional filename timestamp: -YYYYMMDD-HHMM suffix
  • Automatic watermark text timestamp: appended to user text
  • Format: YYYY-MM-DD HH:MM:SS
• Extension Preservation: Maintains original file extension (or upgrades format)
• Metadata Options:
  • Strip original metadata (default for privacy)
  • Preserve original EXIF/metadata (optional)
  • Inject SecureMark identification in metadata

Anti-AI Removal Protection

• Purpose: Disrupt AI model training and prevent automated watermark erasure.
• Level: 0-100% intensity slider.
• Implementation:
  • Position Jitter: Random ±10px offset per watermark (scaled by level).
  • Angle Jitter: Random ±15° rotation variation (scaled by level).
  • Alpha Jitter: Random ±40 alpha channel variation per pixel (scaled by level).
  • Non-deterministic: Different output each time at same settings.

AI Cloaking (Adversarial)

• Purpose: Protect the artistic style of images and block OCR (text recognition) by scraping bots.
• Mechanism: Injects high-frequency noise in the DCT domain.
• Effect: Invisible to human eyes, but causes AI models to misidentify content or fail to extract text data.

PDF Handling

• Vector Engine: Direct PDF graphics layer insertion (preserves text/vector quality)
• Raster Fallback: Automatic page-by-page image conversion for corrupted PDFs
• Rasterization Option: Force rasterization for maximum security (converts to images)
• Page Iteration: Supports multi-page PDF watermarking
• Preview Generation: First page rendered at 72 DPI for preview display

Platform Integration

Android Sharing Integration

SecureMark registers as a share target for image and PDF files:

Intent Filters:

• ACTION_SEND - Single file sharing
  • Accepts: image/* (all image types)
  • Accepts: application/pdf
• ACTION_SEND_MULTIPLE - Multiple file sharing
  • Accepts: image/* (batch image processing)

File Handling:

• Content URI Resolution: Automatic conversion from content:// URIs to file paths
• Extension Preservation: MIME type detection and filename extension recovery
• Supported Formats:
  • image/jpeg → .jpg
  • image/png → .png
  • image/webp → .webp
  • image/heic → .heic
  • image/heif → .heif
  • application/pdf → .pdf
• Cache Directory: Temporary files copied to app cache with unique naming
• Intent History Filter: Prevents re-processing of historical intents

Platform Channel:

• Channel Name: secure_mark/sharing
• Methods:
  • getSharedFiles(): Returns list of file paths from share intent
  • clearSharedFiles(): Clears shared file queue
  • onSharedFilesReceived(count): Callback to Flutter when files arrive

Workflow:

1. User selects "Share to SecureMark" from another app
2. Android system launches SecureMark with SEND/SEND_MULTIPLE intent
3. MainActivity extracts URIs and converts to accessible file paths
4. Files are cached in app-specific directory with original extensions
5. Flutter layer is notified via method channel callback
6. Files are automatically loaded into the watermarking interface

iOS Integration

• share_plus Plugin: Cross-platform sharing API
• File Export: Share processed files to iOS Share Sheet
• Permissions: Photo library access for image picker integration

Desktop Support

• Drag & Drop: desktop_drop package for file dragging (Windows, macOS, Linux)
• File Selector: Native file picker dialogs on all desktop platforms
• Path Provider: Cross-platform temporary and document directory access

Supported Input Formats

• PNG - Lossless compression
• JPG / JPEG - With adjustable quality control
• WebP - Modern format with quality control
• HEIC / HEIF - Native support for modern mobile photo formats
• PDF - Vector or rasterized (flattened) processing

Development & Testing

Install dependencies:

flutter pub get

Run the application:

flutter run

Run security & steganography tests:

flutter test test/unit/steganography_test.dart

Project Goal

SecureMark focuses on providing verifiable proof of ownership and secure distribution for shared documents. It bridges the gap between simple visual watermarking and advanced forensic steganography.

License

This project is licensed under the GNU General Public License v3.0. See LICENSE for details.