No Training Animal Detection Using Vision Language Model

This project introduces a 'zero-shot' approach to wildlife monitoring, allowing users to detect and identify animals in images without any prior training on specific species datasets.

The pipeline integrates GroundingDINO to locate objects based on text prompts (e.g., 'animal') and then passes cropped detections to LLaVA (Large Language-and-Vision Assistant) for refined classification and descriptive analysis.

Designed for flexibility and rapid deployment, the system can adapt to new environments and rare species instantly by simply modifying the text prompts, bypassing the traditional bottlenecks of computer vision workflows.