Computer Vision

knowledge skill

Detects and classifies objects, faces, or text within an image file. Outputs a structured list of detected entities with bounding box coordinates and confidence scores.

Worked examples

• {"input": {"image": "receipt.jpg", "task": "text_extraction"}, "output": {"entities": [{"label": "store_name", "confidence": 0.98, "bbox": [50, 30, 200, 45]}, {"label": "total_amount", "confidence": 0.95, "bbox": [300, 120, 380, 135]}], "image_metadata": {"width": 800, "height": 600, "format": "JPEG"}}}
• {"input": {"image": "street_scene.png", "task": "object_detection"}, "output": {"entities": [{"label": "car", "confidence": 0.92, "bbox": [120, 200, 250, 280]}, {"label": "person", "confidence": 0.88, "bbox": [300, 180, 340, 320]}], "image_metadata": {"width": 1280, "height": 720, "format": "PNG"}}}

Input

• image: file (PNG/JPEG)
• task: string (object_detection|face_recognition|text_extraction)

Output

• entities: array of {label: string, confidence: float, bbox: [x1, y1, x2, y2]}
• image_metadata: {width: int, height: int, format: string}

Details

• Skill type: knowledge skill
• Safety level: safe_public_research
• Version: 1.0.0