Upload Images

Images are the primary asset type for computer vision tasks in Datature Vi. This guide covers everything you need to know about uploading images, including supported formats, file specifications, and best practices.

📖

Main upload guide

For general information about uploading assets (web interface, SDK, tracking progress), see Upload Assets.

---

Supported image formats

Datature Vi supports a wide range of image formats:

💡

Recommended formats

For best results, use JPEG or PNG formats. These provide the best balance of quality, file size, and compatibility across all platforms.

📘

Multi-frame formats

For formats that support multiple frames or subfiles (GIF, AVIF, WebP, TIFF), only the first frame or subfile will be imported as the asset.

---

File size and specifications

Size limits

• Maximum file size: 256 MB per image
• Recommended size: Under 10 MB for optimal performance and faster processing

Image dimensions

• No strict dimension limits — Images of any resolution are supported
• Recommended: Keep images at resolutions appropriate for your use case
• Very large images (e.g., 10,000+ pixels) may take longer to process and render

Quality considerations

• Images are stored in their original format and quality
• No automatic compression or resizing is applied during upload
• Consider pre-optimizing large images to improve upload speed and platform performance

---

Best practices

Image preparation

File optimization:

• Optimize images before upload to reduce file size without sacrificing quality
• Use image compression tools to reduce file sizes while maintaining visual quality
• Remove unnecessary metadata (EXIF data) if file size is a concern

Format selection:

• Use JPEG for photographs and images with many colors
• Use PNG for images with transparency, text, or sharp edges
• Maintain consistent formats within a dataset when possible

File organization:

• Use unique filenames to avoid accidentally overwriting existing assets
• Organize images in folders by category for easier batch uploads
• Remove corrupt or invalid images before uploading

Upload strategy

Batch uploads:

• Upload in batches of 10-50 images at a time for better reliability
• For datasets with 1,000+ images, consider using the Vi SDK for programmatic uploads

Quality checks:

• Verify image quality before upload
• Ensure images are properly oriented
• Check that images are in focus and properly exposed for your use case

Duplicate handling:

• Be aware that uploading a file with the same name will replace the existing asset
• Any annotations linked to the replaced asset will be removed
• Use unique filenames or the SDK's duplicate handling options to prevent data loss

---

Image-specific troubleshooting

Image appears rotated or incorrectly oriented

Possible causes:

• EXIF orientation metadata not being respected
• Image was rotated in the source application but not saved correctly

Solutions:

• Open the image in an image editor and rotate it to the correct orientation
• Save the image with the rotation applied to the actual pixel data
• Re-upload the corrected image

Multi-frame GIF or animated image shows only first frame

Expected behavior:

This is the intended behavior. Datature Vi imports only the first frame from multi-frame formats (GIF, AVIF, WebP).

Solutions:

• If you need to annotate multiple frames, extract individual frames using an image tool
• Upload each frame as a separate image
• For video-like sequences, consider using video uploads instead

HEIC images not displaying correctly

Possible causes:

• Browser compatibility issues with HEIC format
• Corrupt HEIC file

Solutions:

• Convert HEIC images to JPEG or PNG for better compatibility
• Use image conversion tools or export from the source application
• Verify the HEIC file opens correctly in other applications

Large images taking too long to upload

Possible causes:

• Image file size is too large (high resolution or uncompressed)
• Slow internet connection

Solutions:

• Compress images using image optimization tools before upload
• Reduce image resolution if appropriate for your use case
• Upload in smaller batches to improve reliability
• Consider using the Vi SDK for large uploads

---

Common questions

What is the maximum image resolution supported?

There is no strict resolution limit, but very large images (10,000+ pixels) may experience slower upload and processing times. For optimal performance, keep images at resolutions appropriate for your computer vision tasks.

Are images compressed during upload?

No, images are stored in their original format and quality. No automatic compression or resizing is applied during upload or storage.

Can I upload images with transparency?

Yes, PNG images with transparency (alpha channel) are fully supported. The transparency will be preserved in the uploaded asset.

What happens to EXIF metadata?

EXIF metadata is preserved during upload. However, for certain display purposes, EXIF orientation tags may not always be automatically applied.

Can I upload RAW image formats?

RAW formats (CR2, NEF, ARW, etc.) are not currently supported. Convert RAW images to JPEG, PNG, or TIFF before uploading.

---

Next steps

After uploading images, you can:

• Upload annotations — Import existing annotations if you have them
• Annotate data — Add labels manually using the visual annotator
• Train a model — Start training once you have annotations
• View dataset insights — Analyze your dataset statistics

---

Related resources

• Upload Assets — Main guide for uploading assets
• Upload Videos — Video-specific upload guide
• Vi SDK Assets API — Complete SDK reference for asset operations
• Create a dataset — Learn how to create datasets