Spectrum is a semi-universal colorbot that uses C# and OpenCV for image processing.
- Extensive configuration file
- Dataset collection
- Multiple mouse movement paths
On first run, a default configuration file is created.
ImageWidth: 1 - ScreenSize (Default: 640)
ImageHeight: 1 - ScreenSize (Default: 640)
YOffsetPercent: 0.0 - 1 (Default: 0.8)
XOffsetPercent: 0.0 - 1 (Default: 0.5)
EnableAim: true or false (Default: true)
ClosestToMouse: true or false (Default: true)
Keybind: 1 - 254 (Default: 6), these are virtual key codes
Sensitivity: 0.1 - 2 (Default: 0.5)
AimMovementType: CubicBezier, Linear & Adaptive (Default: CubicBezier)
ShowDetectionWindow: true or false (Default: true)
CollectData: true or false (Default: false)
AutoLabel: true or false (Default: false), this is dependent on CollectData
BackgroundImageInterval: 1 - 100 (Default: 10)
UpperHSV
Val0: 0 - 255 (Default: 150) HueVal1: 0 - 255 (Default: 255) SaturationVal2: 0 - 255 (Default: 229) ValueVal3: 0 (Default: 0) ?
LowerHSV
Val0: 0 - 255 (Default: 150) HueVal1: 0 - 255 (Default: 255) SaturationVal2: 0 - 255 (Default: 229) ValueVal3: 0 (Default: 0) ?
When CollectData and AutoLabel are enabled, Spectrum automatically generates:
- Images: Captured screenshots saved to
dataset/images - Labels: YOLO format annotation files saved to
dataset/labels - Background Images: Screenshots without any labels, helps training balance This feature is very useful for creating training datasets for machine learning models.
See LICENSE.txt for license information.
This software is intended for educational and research purposes. Users are responsible for ensuring compliance with applicable laws and terms of service when using this software. The developers assume no responsibility for misuse of this application.
For issues, questions, or feature requests, please open an issue on the project repository.