One of the things I like to do when I go on vacation is taking panoramic photos. Since my camera doesn't take panoramic photos, I have to use software to stitch images together. One of the best image stitchers I have found yet is Image Composite Editor 2 from Microsoft Research Computational Photography Group.

Image Composite Editor (ICE) can take photos taken from a single location that overlap and create stunning panoramic images from them. Using techniques from the field of computer vision, ICE scans the images for similarities to each other and then estimates the vantage point for each photo. ICE can even create panoramic images from videos too.

Sample ICE 2 panoramic image without autocomplete

Sample ICE 2 panoramic image with autocomplete

Microsoft Research Computational Photography Group recently released a new version of ICE. The following is a quote from the ICE website of changes included in ICE 2:

• Redesigned user interface
  ICE has a new look that makes all the features more understandable and easier to use.
• Automatic image completion
  ICE can now fill in any missing pixels around the edges of your panorama, making a smooth boundary even in cases where you missed a shot.
• Improved workflow
  ICE now guides you through the steps required to make a great panorama. And you can back up a step to change settings, then see the effects of those changes without having to start over from scratch.
• Full-resolution preview
  Panorama previews are no longer limited by your screen resolution. ICE now allows you to zoom in to see every detail of your stitched panorama, no matter how big, before you export.
• Built-in Photosynth uploader
  ICE no longer requires the installation of a separate application to upload a panorama to the Photosynth web site. You can now upload and share your panorama directly from within ICE.

ICE will work on 32-bit or 64-bit versions of Windows Vista SP2, Windows 7, Windows 8.1, and Windows 10. For more information on Image Composite Editor from Microsoft Research Computational Photography Group, follow the links below:

ICE Homepage
ICE Support Forums

As many of you know, when I'm not working on computers, I like to work with photos and videos. And some of the coolest software I've found for doing panoramic images has come from Microsoft Research. I have written about Microsoft Research before, Microsoft Image Composite Editor, and HD View, and they have recently released a new project for video called Cliplets.

View of the main screen inside of Microsoft Research Cliplets.

With Cliplets, you can combine static and dynamic elements from a single video to create some pretty cool effects. It works by isolating different elements on individual layers. Each layer has it's own action (still, loop, mirror, or play) and timeline. Just open a compatible video, and Cliplets will ask what segment you would like to use. The maximum amount of time that Cliplets can work with is only ten (10) seconds.

This video is a sample of what Microsoft Research Cliplets can do. The palm tree on the left has been frozen, while the palm tree on the right sways is the wind. Also, look for the bird is flying through.

When you are all done editing, you can export your final Cliplet in three different formats, animated GIF (*.GIF), MPEG-4 video (*.MP4), or Windows Media Video (*.WMA). Cliplets works with 32-bit and 64-bit versions of Windows 7. Here's a quote from their site:

Cliplets: Juxtaposing Still and Dynamic Imagery

What Are Cliplets?

Microsoft Research Cliplets is an interactive app that gives users the power to create "Cliplets" -- a type of imagery that sits between stills and video, including imagery such as video textures and "cinemagraphs". The app provides a simple, yet expressive way to mix static and dynamic elements from a video clip.

About Cliplets

Cliplets is a research project from Microsoft Research. This project explores a form of visual media that juxtaposes still image and video segments, both spatially and temporally, to expressively abstract a moment. The tension between static and dynamic elements in a cliplet reinforces both aspects, strongly focusing the viewer's attention or conveying a narrative. We develop a set of idioms, essentially spatiotemporal mappings, that characterize these cliplet elements, and use these idioms in an interactive system to quickly compose a cliplet from ordinary handheld video. A key challenge is to avoid seam artifacts by maintaining spatiotemporal continuity in the cliplet composition. We address this using several algorithms from computer graphics and vision.

For more information on Microsoft Research Cliplets, follow the links below:

Microsoft Research Cliplets

A while back. I wrote an article about creating panoramic images with Microsoft Image Composite Editor (ICE) and Photosynth. With ICE and Photosynth, you can create panoramic images and then upload them to the Photosynth website. But what if you wanted to upload them to your website? That's where HD View comes into the picture.

Microsoft HD View is a new viewer to display large images (gigapixels) across the internet. It allows you to create panoramic images for web pages that do not take a lot of bandwidth. When someone views the image, they are only downloading what is required for that view. It is only when they pan or zoom that more of the image is downloaded. HD View / Deep Zoom is now included in ICE, also as a stand-alone command-line utility, and (here's the best), a Photoshop plug-in.

HD View / Deep Zoom export in Microsoft Image Composite Editor

HD View / Deep Zoom export in Photoshop plug-in

So if you're a Photoshop user, you can now export those panoramic images that Photoshop can create directly to HD View / Deep Zoom format. The Photoshop plug-in is available in both 32-bit and 64-bit. Here's a quote from the HD View website:

About HD View
HD View is a new viewer developed by Microsoft Research's Interactive Visual Media group to aid in the display and interaction with very large images. The HDView development team included Johannes Kopf, Matt Uyttendaele, Howard Good, and Michael Cohen along with Jonathan Fay of the Next Media group.

Recent advances in camera and sensor technology and software for stitching images together has led to the creation of images containing billions of pixels (gigapixels). These images are often panoramic, that is, they cover very wide fields of view. Since monitors typically contain only one to two million pixels, it is only possible to actually see 1/1000th of such image data at once. Also, viewing very wide fields of view require unwrapping of an image projected onto a curved surface (think of a map of the world) which can cause distortions.

HD View leverages current graphics hardware to allow smooth panning and zooming as well as the viewing transformation described below.

HD View was developed with a number of goals in mind. It should:

• allow smooth panning and zooming on large images,
• only download enough data to create the current view (and possibly look ahead to the next), and
• always display the current field of view with an appropriate projection. This means that when zoomed way in you should be presented with a standard perspective projection providing a sense of immersion, and when zoomed out you experience a curved projection so that get a full overview of the scene. In between the projection should smoothly transition.
• Finally, it should be easy to create your own HD View content and present it to the world via the web.

The HD View plug-in currently supports all major browsers on the Windows platform. The first time that you visit a page with HD View content you will be prompted to install the HD View plug-in.

For more information on HD View / Deep Zoom, follow the links below:

Microsoft Deep Zoom

HD View Utilities (32 bit)

HD View Utilities (64 bit)