User:Yd192/sandbox

Introduction
Here we introduce the method proposed in Exploring Video Streams using Slit-Tear Visualizations .We in some scenarios use cameras to record a fixed scene for later review, as in security or data capture situations.We define this type of video as stationary video scenes, where the footage is typically captured by a strategically positioned fixed-mount camera. Such scenes typically comprise a fixed background and one or more objects that change or move within that scene.slit-tears is introduced here which support the rapid exploration of stationary video scenes for events and patterns of interest.

Related work
Several different approaches for visualizing video data as a timeline that have evolved from slit scanning are list as following: Most non-linear editors offer a visual timeline as the primary way for editors to view and compose video sequences. A panoramic film strip is passed rapidly across a single vertical slit, which exposes film to only a narrow slit from a scene (indeed this is how most digital scanners work). Objects moving in a stationary scene over time are seen as motion captured over space. Sometimes known as video slicing,achieves a similar effect. Here, a slit is placed over a video frame,and the pixels of successive video frames under the slit are captured and added to a composite image. We can also consider a video clip as a volume,where successive video frames are stacked atop one another. i.e., as a video cube. Chung et al. illustrate several methods of rendering this cube, and show how rendering successive frame differences can be a compelling summarization of the video data. Fels et al. provide another visualization by slicing through this cube using a cut plane thereby crossing both time and space. Slit-scanning can be viewed a subset of this method, as it realizes the specific case of an intersecting plane being placed perpendicular to the face of the cube. Video volumes are more general, as different effects can be achieved by using other geometric slicing shapes and other slicing positions.
 * Traditional timelines
 * Slit scanning in photography
 * Slit scanning in video
 * Video volumes

Slit-tears
We have already seen how slit scan photography (and later slit scan video) captures a linear—usually vertical— slice of a frame’s area, and portrays these as instances in time in a visualization. Yet, while photography is technically limited to a single linear slit, there is no need to impose this arbitrary limitation to digital video. Instead, we can capture a moment in time as multiple slit-tears in the video scene that are concatenated together to form a single column in the visualization. With tear-based video slicing, we first allow the end user to draw multiple slits atop a frame’s surface. Each slit is an arbitrary stroke—a straight line, curve, or scribble. For each frame, the system then captures the pixels under each slit in the order and direction that each stroke was drawn, and aligns these pixels into a single vertical column. It then appends this pixel column to the right side of the visualization. We have created two systems that realize the slit-tears method; their combined capabilities are described below. The system works with both live and previously-captured video (AVI files). For live video, the visualization updates itself with incoming video frames, generating new columns corresponding to the slit-tears. For previously-captured video, there are two options. One can treat it like live video, where tears and updates for subsequent parts of the visualization are done as the video plays. Alternately, one can update the entire visualization—past,present and future frames—to reflect the slit-tears. For previously captured video, the user can also select different playback speeds and the level of granularity of playback (the skip frames slider). When frames are skipped, details may be lost. However, the visualization then gives the viewer a broader picture of when and where events happened over longer periods of time.
 * The method
 * The system

Event-level analysis
Slit-tears can visualize and emphasize events of interest. More generally, events of interest can be problematic to see in conventional replay of video when the image quality may be poor, events may be very brief or spatially small, and patterns over time may be hard to detect. Slit-tears are a technique that helps to overcome these difficulties. Slit-tears reveal changes as they occur in a region on a static background. Change highlighting works even if images are blurry, pixelated and/or low-contrast. It will also work over noisy video, as motion tends to produce regular vs. random patterns. Some events of interest may be spatially ‘small’, affecting only a modest number of pixels in the scene. These can be easy to miss or to decipher when replaying a long video. Even if we are expecting an event over just a few pixels, we can make that event highly salient simply by creating several slit-tears over that area; this enlarges the event’s appearance in each column. In a long video scene, the timeline is a series of fairly unbroken horizontal lines; however, when objects do pass through the tears, they appear as an intrusion in the timeline.
 * Events are readily seen even in low-fidelity video.
 * Spatially “small” and/or poor fidelity events are exaggerated.
 * Brief events are made extremely salient by virtue of how the timeline is constructed.

Pattern-level analysis
One of the strengths of slit-tear visualization is that it allows us to easily see not only individual events, but patterns in the scene. By patterns, we refer to how events relate to one another over time. Because the tears traverse space, and the timeline traverses time, temporal patterns of movement and behavior are visualized spatially. As well, events can be more easily correlated when the analyzer strategically places slit-tears to juxtapose them in the timeline. Movement or events that reoccur through a tear are strikingly easy to see since they appear in the timeline as a repeated intrusion on the scene. When multiple similar events are occurring, the analyzer can juxtapose them in the visualization by the ordering of the tears. Slit-tears of different events can be juxtaposed to see whether correlational relationships exist between objects, movement, or patterns in the video scene. With strategic placement of slit-tears, the directionality and comparative velocity of objects moving about the scene can be easily ascertained.
 * Rhythms and periodicity over a tear are easy to see.
 * Similar individual events can be compared as a category.
 * Different individual events can be correlated.
 * Directionality and velocity can be easily ascertained.

Advantages

 * Slit-tears allows for exploratory video analysis.
 * Slit-tears is a generalizable video-analysis technique.
 * Slit-tears is grounded in the actual data.

Limitations

 * It only works with stationary video.
 * Appropriate camera placement is crucial.
 * The analyzer must be immersed in the data.