FNBubbles420 Org - Assistive Aim-Guide

Assistive Accessibility Game Tool 🎯

Welcome to the Assistive Accessibility Game Tool project! This tool supports individuals with disabilities by detecting and aiming at human targets within a game window using ONNX-based object detection.

---

Discord Support

Join our Discord channel Assistive AimGuide for assistance, support, or to share your experience.

HAVING ISSUES ??

• JOIN THE DISCORD https://github.com/FNBUBBLES420-ORG/Assistive-Accessibility-Game-Tool/blob/main/having-issues.md

---

🔄 Recent Updates

• Configurable Game Window Title: Added gameWindowTitle in config.json for CPS focus detection.
• Serial Toggle: Unified flag to enableSerial; removed deprecated serial key; launcher UI now edits enableSerial only.
• Active Window CPS: CPS counter now measures only when the configured game window is in focus.
• Mask Toggle Fix: Proper boolean coercion of useMask strings to enable/disable masking.
• Movement Amp Decimals: Support for decimal aaMovementAmp values (e.g., 1.1, 1.2); update config.json accordingly. If decimals aren’t applied correctly, join the Discord for a quick fix.
• Style Constant Fix: Corrected RESET_ALL typo in colorama usage.

---

🧠 What It Does

• Captures game feed via dxcam or bettercam.
• Runs frames through an ONNX model (e.g. YOLOv5) for human detection.
• Applies non-max suppression to filter detections.
• Smoothly moves the mouse toward the closest target.
• Supports headshot mode and center-of-screen targeting.
• Optional masking of screen regions to ignore overlays.
• Draws bounding boxes with OpenCV directly on the feed.
• Displays CPS in the console when the game window is active.

---

🚀 Getting Started

1. Install Python

Download and install Python 3.11.9 via:

python3119.bat

Note: You can set aaMovementAmp to decimal values such as 1.1, 1.2, 1.3, etc. Update the config.json entry accordingly under aaMovementAmp. If the decimal value isn’t applied properly, join our Discord for a quick patch.

2. Auto Setup

Unzip autosetup.zip and run the auto-setup executable to install all required dependencies automatically:

autosetup.exe

2. Install Requirements

pip install -r requirements.txt

---

🚀 NVIDIA Setup Instructions

1. Install CUDA 11.8
   • Download from the NVIDIA archive: https://developer.nvidia.com/cuda-11-8-0-download-archive
   • Run the .exe installer and ensure you select:
     • CUDA Toolkit
     • CUDA Samples
     • Visual Studio Integration (optional)
     • CUDA Documentation (optional)
2. Reboot your system after the CUDA installation.
3. Install cuDNN 8.9.6
   • Download from NVIDIA: https://developer.nvidia.com/downloads/compute/cudnn/secure/8.9.6/local_installers/
   • Unzip and copy all files into:

     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8
     

4. Install TensorRT
   • Download from NVIDIA: https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/secure/
   • Extract and merge contents into your CUDA directory (same path as above).
   • Then install the matching Python wheel, for example:

     pip install "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\python\tensorrt-8.6.1-cp311-none-win_amd64.whl"

5. Install CuPy for CUDA

   pip install cupy-cuda11x

6. Add Environment Variables Add these paths to your system environment variables:

   C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin
   C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\libnvvp
   C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\lib
   

7. Verify GPU Availability

   import torch
   print(torch.cuda.is_available())
   print(torch.cuda.get_device_name(0))

8. Install Visual Studio 2022
   • Install Community Edition with Desktop development with C++ workload.

---

🔦 AMD Setup Instructions

1. Install AMD GPU Drivers
   • Download and install the latest AMD Adrenalin drivers from https://www.amd.com/en/support.
2. Install Visual Studio 2022
   • Community Edition with Desktop development with C++ workload selected.
3. Install DirectX End-User Runtime
   • Download from: https://www.microsoft.com/en-us/download/details.aspx?id=35
4. Install ONNX Runtime with DirectML

   pip install onnxruntime-directml==1.17.0

5. Verify DirectML Provider

   import onnxruntime as ort
   print(ort.get_available_providers())  # should include 'DmlExecutionProvider'

---

3. Edit and Save Your Configuration

Launch launcher.bat to open the GUI. Adjust settings and click Save:

• gameWindowTitle: e.g. "Fortnite"
• enableSerial: true or false
• useMask, cpsDisplay, visuals, etc.
• fallbackSerialPort, baudRate, onnxProvider

Example config.json:

{
  "screenShotHeight": 320,
  "screenShotWidth": 320,
  "aaMovementAmp": 1,
  "useMask": "false",
  "maskHeight": 240,
  "maskWidth": 140,
  "maskSide": "left",
  "aaQuitKey": "8",
  "aaActivationKey": "CapsLock",
  "confidence": 0.72,
  "headshot_mode": "true",
  "cpsDisplay": "true",
  "visuals": "true",
  "gameWindowTitle": "Fortnite",
  "centerOfScreen": "true",
  "model_path": "onnx-models/v5_mini_royale.onnx",
  "enableSerial": false,
  "fallbackSerialPort": "COM3",
  "baudRate": "115200",
  "onnxProvider": "CUDAExecutionProvider"
}

---

4. Run the Main Application

Use launcher.bat to start the application. In the GUI:

1. Adjust your settings as needed (sliders, dropdowns, text inputs).
2. Settings will auto-save as you adjust them—no separate “Save” button needed.
3. Click Launch Main Onnx to begin the assistive aim tool.

---

✅ Required Python Packages (requirements.txt)

# torch without CUDA support is for AMD GPUs
torch>==2.7.0
torchvision>==0.22.0
torchaudio>==2.7.0
# Uncomment the following lines if you are using Nvidia GPUs
# torch>==2.7.0+cu118
# torchaudio>==2.7.0+cu118
# torchvision>==0.22.0+cu118
numpy==1.26.4
# uncomment if you have Nvidia gpu below 👇 
# onnxruntime-gpu==1.17.0 # This is correct for this application 
onnxruntime-directml==1.17.0 # This is correct for this application 
opencv-python>==4.8.1.78
comtypes
customtkinter
pandas
# uncomment the following line Below if you are using Nvidia GPUs
# cupy-cuda11x
bettercam
psutil
colorama
ultralytics
PyAutoGUI
PyGetWindow
pywin32
pyyaml
tqdm
matplotlib
seaborn
requests
ipython
dxcam
onnx>==1.17.0
onnx-simplifier
dill
serial
pyserial
pyarmor
# DirectML dependencies for Windows AMD GPUs
# torch-directml

---

⚙️ Supported GPU Providers

• CUDAExecutionProvider – NVIDIA GPUs (CUDA/cuDNN)
• DmlExecutionProvider – AMD GPUs (DirectML)
• CPUExecutionProvider – CPU-only

---

🧩 System Overview

• Capture video from the game window.
• Detect humans in each frame via ONNX model.
• Suppress overlapping boxes.
• Aim the cursor smoothly at targets.
• Display bounding boxes and CPS metrics.

---

❗ Important Notes

This software is an assistive tool for gamers with physical or cognitive impairments. Part of FNBubbles420 Org, a registered 501(c)(3) charity.

---

🙌 Support & Community

Join our Discord: https://discord.gg/9vG777kJYc

© 2025 FNBubbles420 Org. All Rights Reserved.