Constructing a SenseCam visual diary as a media process

Abstract

The SenseCam is a small wearable personal device which automatically captures up to 2,500 images per day. This yields a very large personal collection of images, or in a sense a large visual diary of a person's day. Intelligent techniques are necessary for effective structuring, searching and browsing of this image collection for locating important or significant events in a person's life. In this paper we identify three stages in the process of capturing and structuring SenseCam images and then displaying them to an end user to review. These stages are expressed in terms of the canonical process stages to which they correlate.

Computer Vision for Lifelogging

“…One day's continuous visual lifelogs such as image streams can usually be segmented such as using the technique introduced in [42], into dozens of events which eases the representation and interpretation of everyday engagements. A lifelog event corresponds to a single activity in the wearer's day such as watching TV, commuting, or eating a meal, and the durations vary a lot due to the engagement natures of various activities.…”

Digital Chronofiles of Life Experience

“…1. An example lifelog browser [12] There are many potentially life enriching benefits that could encourage an individual to engage in lifelogging, such as the potential for better self-awareness leading to longer and more active lifespans, new personalised healthcare applications, enhanced methods of learning, increased productivity in the workplace, increased independence, or mobility for people suffering from various memory and cognitive impairments, and new forms of offline and online social interaction [7]. It is considered that as these benefits become apparent, that lifelogging will become a normative activity.…”

“…Many use-cases of lifelogging have been proposed, for example, to assist the lifelogger in tasks such as personal healthcare, memory reminiscence [3], and browsing a digital record of past activities [12]. There is now real potential that we are on the cusp of an era of, what Gordon Bell and Jim Gemmell refer to as total capture [2].…”

Issues in Designing Novel Applications for Emerging Multimedia Technologies

“…The design of a lifelogging photo browser where a large number of a person's lifelog photos from a SenseCam can be reviewed [15] took more than a year during which the strategies to connect between the supporting back-end techniques and their possible front-end manifestation and interaction schemes gradually developed -there was no precedence of such an application and there still is no usage of such an application as it will take many more years for the passive capture device such as SenseCam to be used by the general public. Many of the novel applications incorporating multimedia techniques developed in our group took months for each of them to form any concrete user-interface, because of the lack of understanding in the application areas and the lack of examples to narrow down the design space.…”

“…a vase, a car or a ship) in terms of the object characteristics rather than whole photos [25]; a security video search system where CCTV video footage in a uni- versity campus is captured and indexed by people's contours and later a security staff can trace whereabouts of a suspicious person appearing in multiple camera locations [14]; a museum artefact explorer where a user can upload the photos taken at the museum to match and group the same artefacts even if the photos show different angles or sides, and display information on them (see Figure 1 for a screen shot) [2]; an online photo organiser where the uploaded personal photos are automatically annotated by the people appearing in the photos [23]; a Lifelogging browser where a large number of passively captured photos from a wearable camera are automatically structured by individual events, their relative importance identified and presented in an intriguing comic-book style montage as to help the user review their day [15]; a route finder where a collection of invehicle video footage is automatically indexed and seamlessly interweaved with the vehicle routes on a geographic map [19], and many more. All of these applications incorporated one or more combinations of multimedia techniques that were on-going research topics at the time (and many of them still are), and were more than just "technology demos" in the sense that an extensive Interaction Design effort was expended in developing them and specially designed to support some form of novel activities that conventional applications do not support.…”