Comparing pixel color sampling (Desktop Duplication API / GDI / Direct3D 11 screen capture) versus memory-based entity coordinates:

Pixel Scanning Approach:
- Uses to grab GPU framebuffers in ~1ms.
- Searches a bounding region (e.g. 100x100 pixels around crosshair) for high-contrast enemy outlines (like Purple/Yellow enemy highlights).
- Zero memory reading / zero process interaction.

Memory Coordinate Approach:
- Requires reading 3D position vectors from player structures.
- Projects 3D world space to 2D screen coordinates with matrix math.
- Provides 360-degree awareness and distance scaling.

What are your thoughts on performance tradeoffs between Desktop Duplication and Direct3D overlay pipelines?

DXGI Desktop Duplication is surprisingly efficient on modern GPUs when using hardware format conversion (). Filtering by HSV color space (Hue, Saturation, Value) instead of raw RGB makes yellow/purple highlight detection very accurate.

Desktop Duplication avoids any in-process hooks entirely. For internal tools, DirectX Present hooks offer higher frame rates, but for purely external utilities, Desktop Duplication is rock solid.