Online Computational Stereo Vision

IVS Lab

Minh Nguyen
Computer Science Department
The University of Auckland

[Server is ACTIVE, reported at 2026-05-19 06:36:46] [Session: None]

Online Computational Stereo Vision - New Frontend Under Development

Choose types and upload stereo images, select from the following input types:

Upload left/right images

Upload combined image

Upload anaglyph image

Upload .MPO .PNS .JPS

Upload stereogram image

Required image rectification:

(*) Horizontal alignment of the image pair

Stereo matching algorithm:

Choose one

Force Dispary Range:

(*) Force depthmap to result N depth levels

minimum disparity:

(*) Negative value is allowed

maximum disparity:

(*) Positive value only

Disparity scale:

(*) Positive value only and 1 is for auto scale

Smoothing value:

(*) Positive value only and less than 50, ie: 1 2 3...

Resize large images to:

pixels (the larger, the slower)

Average waiting time: ?? ± ?? seconds excluding upload and download time!
or Reset

File allowed: JPG, JPEG, PNG, PPM, MPO, large images will be automatically resized to the above width.
If you only change parameters, click Submit again to re-calculate the result.
Click here to access the gallery page, or check our demo page.

Below is the previously processed dataset, hover over to see its 3D scene.
Processed parameters: width 1021 px, height 576 px, min disparity 3 px, max disparity 73 px, disparity map is re-scaled to 256 levels.
Click to view it in an external page?


Open in WebVR
Downloads: Processed inputs: Resized left image, Resized right image, Left rectified image, Right rectified image, Disparity-map/Depth-map: Rectified Left Disparity Map, Left Image Disparity Map, Grey Depth Map, Colour Depth Map, Shared Combined Results Optical views: Anaglyph, Cross Eyed Side-by-side, Parallel Eyed Side-by-side, AutoStereogram Alias Wavefront’s OBJ files: Model.obj, Model.mtl, Model.jpg (Notice: you need to download all these three files to have a 3D model with texture, use GLC Player to view) Other Parameters: Fundamental Matrix, Disparity Range Parameters, Left/Right Correspondences,