Back to Jue Wang's homepage

Mobile Collaborative Photography

last modified: 7/9/2015


Photography has been largely a one person’s task since the invention of the camera in the 18th century. While setting up the scene may require group effort, the camera has always been designed for a single photographer who controls all the key factors in the capturing process, from scene composition and camera setting to the shutter release time. Thus, despite the rapid advances on camera hardware in recent years, the basic workflow of taking a picture, i.e. a single camera controlled
by one photographer, largely remains unchanged.

We ask ourselves this quesiton: can we make photography a collaborative, social, and fun experience?

Our idea is to allowing a group of users to work as a team to capture a single photo or video. By doing so, we aim to introduce an entirely new social photo capturing experience. This trend is in fact already happening. The recent development of mobile
device cameras enables consumers to take high quality pictures anywhere at anytime, and further allows them to be
shared on social networks right after being captured, transforming photography into an essential component in our daily
social lives. The creative community also started to explore new ways to allow a group of users to collaboratively develop
a photo story using photos from different users. However, all these social interactions happen after the photos have
been taken. In contrast, our system enables teamwork and user interactions before a photo is taken, when the users are
at the same physical location together.

With this methodology in mind, we have developed the following two systems that try to address two specific applications.

CamSwarm: creating bullet-time effect video using phones

In this project, we develop a new App called “CamSwarm”, and demonstrate novel imaging experience inspired by those provided by professional CamArrays. In contrast to traditional CamArrays needing hundreds of cameras and complex calibration, our system allows nearby smartphones to quickly form a small-scale camera array, and synchronizes
all devices to allow them to capture at the same time. To help users position and direct their cameras, we provide a realtime
interface to guide users to move, so that the cameras are properly spaced along a circle that centers at the target object,
and are all pointed to it. Using our system, an instantaneous, wireless smartphone camera array can be set up very quickly
(e.g. within a minute).      

                                       Top: professional camera array. Bottom: our system in action.

Using our system, a group of novice users can work together to capture bullet-time videos like shown bellow. These are still frames extracted from the final rendered video clips. Yes there are artifacts, but remember they don't cost you anything to create (assuming our future offical app is free ;)), and the whole process is just FUN!

    Two examples of the created bullet-time videos using this app. Now your amature jump becomes much cooler!

 You can find more details about the system in our research paper and video:

Yan Wang, Jue Wang, Shih-Fu Chang.  CamSwarm: Instantaneous Smartphone Camera Arrays for Collaborative Photography. arXiv:1507.01148 , July 4, 2015. 

PanoSwarm: creating panoramas of dynamic scenes together

With a similar concept, we develop another system allows multiple regular mobile devices to dynamically form a team, and work collaboratively in a synchronized fashion to capture a high quality panorama of a dynamic scene. The video below explains how it works:

Here we compare panoramas captured by the default iPhone app, and our app. We intentionally shot dynamic scenes to highlight the benifit of our system. Bottom line: because all images are catpured at the same time using our system, you would never see ghosting artifacts!


The paper below contains more details about the system:

Yan Wang, Jue Wang, Shih-Fu Chang.  PanoSwarm: Collaborative and Synchronized Multi-Device Panoramic Photography. arXiv:1507.01147, July 4, 2015.