VidCrit: Video-based Asynchronous Video Review
summary
VidCrit is a two-part system: a feedback recording interface that captures speech + webcam + mouse and the reviewer’s scrubbing actions, and a viewing interface that transcribes via rev.com and auto-segments the session into topical comments using pauses, scrubs and ML features. In studies reviewers produce on average 22 vs 10 comments using recording versus text, segmentation outperforms a sentence-baseline, and authors prefer the segmented viewing interface to email for navigating feedback.
AI-written summary · about
chapters (8)
- 0:00Introduction and motivation
- 2:56VidCrit system overview
- 5:06Feedback recording interface
- 7:11Feedback viewing interface
- 9:10Segmentation algorithm details
- 11:38User study results
- 14:09Future work and limitations
- 15:34Q&A session
related talks
19mWhat Happens to Digital Feedback?: Studying the Use of a Feedback Capture Platform by Care ...
14mRescribe: Authoring and Automatically Editing Audio Descriptions
12mCoSight: Exploring Viewer Contributions to Online Video Accessibility Through Descriptive Commenting
29mCode review, do you speak it?
22mThe truth about code reviews
11mNoteIt: A System Converting Instructional Videos to Interactable Notes Through Multimodal Video U...
43mCode Review Tools
30mCode Reviews Using Art Critique Principles
