US Visa Photo Checker Now With TensorFlow Face Detection
We've upgraded our US visa photo checker with TensorFlow-powered face detection and several new validation checks. The tool now catches more issues that cause State Department rejections—before you submit your DS-160 application.
For complete requirements, see the official State Department photo guidelines .
TensorFlow Face Detection
The headline feature: automatic face detection using a TensorFlow neural network model running directly in your browser.
What It Checks
Head Size Validation — The State Department requires your head to occupy 50-69% of the image height. The checker now measures this automatically and tells you if your head is too large or too small in the frame.
Face Centering — Off-center faces trigger rejections. The tool detects your face position and flags photos where you're not properly centered.
Eye Detection — Closed eyes are an automatic rejection. The checker identifies when eyes appear closed and warns you to use a different photo.
Profile Detection — Visa photos require a front-facing view. The tool detects sideways or profile shots and prompts you to use a photo taken straight-on.
No Face / Multiple Faces — The checker identifies when no face is detected (blurry photos, obstructions) or when multiple people appear in the frame.
Why This Matters
Head positioning errors cause over 56% of photo rejections. Previously, applicants had to estimate whether their head size and position met requirements. Now you get immediate, precise feedback before uploading to the DS-160 form.
Background Color Validation
The tool now analyzes your photo's background to verify it meets the white or off-white requirement.
It examines edge pixels around the frame to detect:
- Colored backgrounds (blue, gray, patterned)
- Gradients or uneven lighting
- Shadows cast on the background
No more guessing whether your background is "white enough"—you'll know before you submit.
Photo Quality Analysis
Beyond face detection, the checker now evaluates overall photo quality:
Brightness Detection — Flags photos that are too dark (underexposed) or too bright (overexposed). Both conditions cause rejections because facial features become difficult to verify.
Contrast Validation — Low contrast images where facial features blend together get flagged for resubmission.
Progressive JPEG Warning — The DS-160 system can have issues with progressive JPEGs. The checker identifies this format and recommends converting to a standard JPEG.
Photo Age Detection — Using EXIF metadata, the tool checks when your photo was taken. If it's more than 6 months old, you'll see a warning since the State Department requires recent photos.
Improved Technical Validation
We've refined the existing dimension and format checks:
Tolerant Aspect Ratio — The checker now allows slight variations from the perfect 1:1 ratio (like 600x602 pixels) that don't affect acceptance. This reduces false warnings for photos that will actually pass.
Minimum File Size — Very small files often indicate corrupted or extremely low-quality images. The checker now flags these before you attempt an upload.
7 Languages Supported
All validation messages and guidance are now available in:
- English
- Spanish
- Russian
- Chinese
- Vietnamese
- Romanian
- Hindi
Select your language from the dropdown and all error messages, instructions, and recommendations appear in your preferred language.
For Immigration Agencies
If you're processing visa applications for clients, these improvements help catch issues earlier in your workflow. The face detection and background validation run instantly in the browser—no uploads to external servers required. This means you can validate client photos at scale without privacy concerns.
Try It Now
The updated photo checker is live. Upload a photo and you'll see the new TensorFlow-powered checks in action.
Have a photo that passed the checker but got rejected by the State Department? Let us know—we're continuously refining the detection thresholds based on real-world feedback.