🎯 RN_AI_cpp — AI Aim Assistant

🚀 Quick Start • 📚 Documentation * 📚Русская версия README

---

✨ Features

🎯 AI Detection	Target detection using neural networks with high accuracy
🎨 Color Detection	Target identification through color-based filtering
📈 Real-time Stats	FPS counter and latency information
🖱️ Aim Simulator	Visualization of target movement prediction
🎛️ ClassTable	Real-time dynamic class management
🔄 Kalman Filter	Smooth movement without aim jitter
⚡ Multiple Backends	DirectML, CUDA+TensorRT, Color Detection

---

🚀 Quick Start

1️⃣ Choose Your Build

🟢 DirectML (Universal)

For: Any GPU (NVIDIA, AMD, Intel, integrated graphics)

✅ Windows 10/11 (x64)
✅ No CUDA required
✅ Recommended for older GPUs

Recommended for:

• GTX 10xx/9xx/7xx series
• AMD Radeon GPU
• Intel Iris/Xe GPU
• Laptops and office PCs

🟡 CUDA + TensorRT (Maximum Performance)

For: Latest generation NVIDIA GPUs

✅ RTX 2000/3000/4000 and newer
✅ GTX 1660
✅ CUDA 12.8 + TensorRT 10.8 (included)
❌ Does not support GTX 10xx/Pascal and older

Features:

• Switch between CUDA+TensorRT and DML in settings
• Maximum FPS and accuracy
• Professional-grade performance Realese — https://mega.nz/file/T8IFHS7I#70_WjY_-3rDZ82U3yKS3meS8mk3bV29_RFrSCQRlFhg

---

⚙️ Configuration & Parameters

📦 ClassTable — Class Management

Dynamic management of target classes with:

• ✅ Add/switch classes in real-time
• ✅ Auto-detection of new classes
• ✅ Configure Y1/Y2 position for each class

---

🎨 Colorbot — Color Detection

Advanced color filtering system:

Parameter	Range	Description
color_erode_iter	0-5	Number of erosion iterations (reduces noise)
color_dilate_iter	0-5	Number of dilation iterations (restores size)
color_min_area	1-1000	Minimum object area
color_target	Yellow/Red/etc	Target color for tracking
tinyArea	1-100	Small element filtering threshold
isOnlyTop	true/false	Consider only top objects
scanError	0-100	Allowed search error (0=precise)

💡 Use: Accurate color-based target selection while ignoring noise

---

🎯 Kalman Filter — Movement Prediction

Smoothing filter for target position prediction:

Parameter	Description
kalman_process_noise	Accounts for random movement changes
kalman_measurement_noise	Accounts for sensor/camera errors
kalman_speed_multiplier_x/y	Speed multiplier per axis
resetThreshold	Filter reinitialization threshold

💡 Result: Smooth aiming without jitter

---

🖥️ Interface & Controls

🎨 ImGui Menu

Menu interface features:

• 🧭 Vertical navbar with custom icons
• 🖼️ Custom background via ui_bg.png
• 🎨 Theming in ui_theme.ini
• ⚙️ Components tab for runtime configuration

📸 Interface Screenshots

Screen Capture	Target Status

🎛️ Overlay Controls

• Overlay Opacity — Transparency (slider or ±)
• UI Scale — Interface scale (± or manual input)
• Window Width/Height — Window size (manual input)
• Resize Handles — Resize window from edges

🎮 Game Overlay — On-Screen Visualization

Information displayed directly on desktop over games and apps:

• 📊 Stats — FPS counter and latency info
• 🎯 Aim Simulator — Aiming prediction visualization
• 🔲 Detection Boxes — Detected target boxes
• 🎨 Class Colors — Auto-coloring (class 0 = green)
• 📝 Text Size — Adjustable in Components → Advanced

---

🔧 Technical Details

📁 File Structure

File	Purpose
config.ini	Main project configuration
ui_theme.ini	UI colors, sizes, and parameters
ui_bg.png	Menu background image (replaceable)
imgui.ini	Window state (local, not committed)

📦 Core Modules

📹 capture/ — Screen capture methods

• DirectX Duplication API — duplication_api_capture
• Windows Runtime capture — winrt_capture
• 📖 OBS Capture — obs_capture

🧠 detector/ — Target detection system

• DirectML detector — dml_detector
• TensorRT detector (NVIDIA) — trt_detector
• Color-based detection — color_detector

🎨 overlay/ — Visual interface

• ImGui implementation — imgui_impl_*
• 2D/3D rendering — rendering

⚡ Input Methods

• WIN32 API — Built-in Windows APIs
  ⚠️ Warning: Don't use in games (instant detection)

• Makcu/Kmbox/KmboxNet — Specialized input devices
  ✅ Recommended for games (low latency)

---

📚 Links & Resources

📖 Documentation

• 🔗 TensorRT Docs
• 🔗 OpenCV Docs
• 🔗 CUDA 12.8
• 🔗 Config

🛠️ Libraries Used

Library	Purpose
ImGui	User Interface
OpenCV	Computer Vision
TensorRT	Neural network inference (NVIDIA)
DirectML	GPU computing (universal)
CppWinRT	Windows Runtime APIs
GLFW	Window management
nlohmann/JSON	JSON processing

💡 Methods & Inspiration

• 🔗 WindMouse Algorithm — Natural mouse movement
• 🔗 KMBOX — Input device integration
• 🐍 RN_AI (Python version)
• 🔀 Original SunOne Aimbot — RN_AI_cpp is a complete fork and rebuild

---

Made with ❤️ for the gaming community