Generate 870k Realistic Synthetic Images

I’m expanding a small-scale computer-vision project and need roughly 870,000 synthetic 2-D images produced in JPEG format. These files will feed the training phase of my model, so visual realism is essential—nothing that looks obviously rendered to the naked eye.

What you’ll start with
• A folder of representative reference shots of semiconductor chips.
• A CSV that spells out the colour palette, precise X-Y measurements and other per-image parameters.

Core expectations
• Build an automated pipeline in Unity (my preferred tool) that ingests the CSV, applies the specified colour accuracy and shape detail, and exports camera-ready JPEGs.
• Use Python or C++ scripts to drive batch generation, manage naming conventions, and verify output counts.
• Keep all images photorealistic and strictly within the provided dimensional tolerances.
• Deliver reproducible project files so I can regenerate or tweak the dataset later.

Acceptance criteria
1. Around 870,000 JPEGs generated, each matching its CSV line (colour, shape detail, size).
2. Spot checks of 500 random samples pass a side-by-side visual inspection against references.
3. Automated script logs confirm zero missing IDs and consistent resolution throughout.

If you already have three or more years creating technical imagery—especially of semiconductor components—and can leverage Unity’s 2-D toolset efficiently, let’s talk timing and milestones.