How to Create Photogrammetry and Gaussian splats from Existing Images and Videos

Polycam lets you upload images or videos on Web or iOS to create detailed 3D photogrammetry models processed in the cloud.

Plan Availability

Before You Upload

Before uploading, you'll want to make sure your images or video have enough coverage and overlap for Polycam to reconstruct an accurate 3D model.

Uploading Images

Take multiple photos of your object or environment from different angles. To get the best results:

• Make sure your photos overlap by 60-80% to allow for proper stitching during processing.
• Depending on the complexity of your subject, you may need anywhere between 20 and 2,000 images.
• Accepted formats: PNG and JPG.

Uploading Video

If you prefer to upload video, Polycam will automatically extract frames for 3D reconstruction. To get the best results:

• Keep your footage steady and make sure it covers the entire subject from all necessary angles.
• Your video must be a minimum of 15 seconds and no longer than 30 minutes or 16 GB.
• Accepted formats: MP4, MOV, AVI and M4V.

Getting Started

The upload and processing flow is the same on Web and iOS. Follow the steps below for your platform to get to the upload screen, then continue with the steps in the next section.

On Web

Go to poly.cam and sign in to your account. On the left-hand side of the page, click the Create button and select Create model from images/video from the drop-down menu.

On iOS

In your Polycam capture library, tap the + icon in the top right-hand corner. From the drop-down menu that appears, select Upload media. On the Create 3D Model page, tap either Pick Photos or Pick Video to open your camera roll and select the files you want to upload.

Upload and Processing Steps

1Upload Your Files

Select or drag and drop your images or video into the upload area. Make sure your files meet the requirements below before uploading.

2Choose Your Processing Mode

Once your files are uploaded, you'll need to select how you'd like Polycam to process them:

• Photogrammetry: Available on all plans. Reconstructs a detailed 3D model from your images or video frames.
• Gaussian Splat: Available on Basic and above. Creates a high-fidelity neural rendering of your scene, following the same workflow as a capture created via Object Mode.

3Select Your Processing Quality

If you selected Photogrammetry mode, choose the level of detail you'd like for your 3D model output:

• Optimized: Provides a good balance of high resolution and usability. Best for game engines and iOS/Mobile.
• Medium: Moderate detail suitable for general use.
• Full: Maximum detail for professional workflows.
• Raw: Unprocessed output for advanced users.
• Custom: Configure your own output parameters.

4Configure Advanced Settings (Optional)

Open Advanced Settings to access additional options that can help improve your results.

Photogrammetry options:

• Smart Keyframing: Allows Polycam's algorithm to automatically select the best non-blurred keyframes from your footage for clearer reconstruction results.
• Isolate object from environment: If you turned or flipped your object during capture, enable this option to help Polycam correctly identify the subject.

Gaussian Splat option:

• Smart Keyframing: Automatically selects the optimal frames from your footage for the best splat quality.

5Start Processing

Once your files and settings are ready, click or tap Process to begin. Polycam will upload and process your capture in the cloud. Depending on the number and size of your files, processing may take anywhere from a few minutes to an hour. You can track the progress of your capture from your capture library.