Why vendor photo upload systems reject your image

After having uploads rejected five times across different orders, I have a pretty complete picture of why automated review systems flag photos. Here is what they are actually checking.

Face detection failure

Vendor upload systems typically run a face detection pass before accepting a photo. If the algorithm does not detect a face confidently — because of a hat, strong side-lighting creating shadows, a slightly turned head, or an unusual angle — the system rejects the upload as unusable for ID production. Even if a human can clearly see your face, the algorithm may not.

The fix: photo straight on, head level, no hat, eyes open and looking at the camera. This is exactly how a passport photo is taken because passport photo requirements are based on the same face-detection requirements.

Background issues

Most vendors require a neutral background. Their systems either detect non-neutral backgrounds automatically or a human reviewer rejects the photo. Cluttered backgrounds, patterned walls, and outdoor backgrounds with foliage or sky all get rejected. Plain white or light grey is safest. Some states require blue — ask the vendor if unsure.

Resolution and file format

Photos below a minimum resolution get rejected with a quality error. HEIC files from iPhones are a common cause of silent upload failures — the file uploads but the backend processing fails because it cannot handle the format. Convert to JPEG before uploading.

Lighting quality

Some systems run an automated lighting quality check — overexposed photos, photos with strong shadows on the face, and photos with warm indoor color casts can all trigger rejection. Even illumination from the front, no harsh directional shadows, neutral color temperature. Natural daylight from a window (indirect, not direct sun) is the most reliable lighting setup for passing automated checks.

Vendor photo requirements are stricter than most social media or government ID photo requirements because the card production workflow is optimized for automated face detection and positioning. Photos that work fine for a passport application sometimes fail vendor upload checks because the cropping algorithm expects a tighter face crop with specific pixel dimensions. Reading the vendor's photo specification page before taking the photo avoids most rejections.