Browsing Sport Content through an Interactive H.264 Streaming Session

Abstract

Abstract-This paper builds on an interactive streaming architecture that supports both user feedback interpretation, and temporal juxtaposition of multiple video bitstreams in a single streaming session. As an original contribution, these functionalities can be exploited to offer improved viewing experience, when accessing football content through individual and potentially bandwidth constrained connections. Starting from a conventional broadcasted content, our system automatically splits the native content into non-overlapping and semantically consistent segments. Each segment is then divided into shots, based on conventional view boundary detection. Shots are finally splitted in small clips. These clips support our browsing capabilities during the whole playback in a temporally consistent way. Multiple versions are automatically created to render each clip. Versioning depends on the view type of the initial shot, and typically corresponds to the generation of zoomed in and spatially or temporally subsampled video streams. Clips are encoded independently so that the server can decide on the fly the version to send as a function of the semantic relevance of the segments (in a user-transparent basis, as inferred from video analysis or metadata) and the interactive user requests. Replaying certain game actions is also offered upon request. The streaming is automatically switched to the requested event. Later, the playback is resumed without any offset. The capabilities of our system rely on the H.264/AVC standard. We use soccer videos to validate our framework in subjective experiments showing the feasibility and relevance of our system.

Resource Allocation for Personalized Video Summarization

“…Instead of using (complex) semantic scene analysis tools, we segment the video based on the monitoring of production actions by analysing the view-structure [37]: We detect replays from producer-specific logos [44], extract shot-boundaries with a detector proposed in [45] to better deal with smooth transitions, and recognize the view-type by using the method in [26]. As in [36], we automatically locate hot-spots by analysing audio signals [46], whose (change of) intensity is correlated to the semantic importance of each video segment.…”

Automatic summarization of broadcasted soccer videos with adaptive fast-forwarding

“…In contrast, event information acts at a global level, and directly impacts the (non)inclusion of segments in the summary. As done in [8], we also detect replays from producer-specific logos [10], extract shot-boundaries with a detector that has been developed in [11] to better deal with smooth transitions, recognize the view-type by using the method in [12], and automatically locate hot-spots by analyzing audio signals [13]. Detailed explanation of those methods are omitted due to page limitation.…”

Real-time CNN-based Segmentation Architecture for Ball Detection in a Single View Setup

“…Locating the ball in team sports is an important complement to player detection [28] and tracking [18], both to feed sport analytics [36] and to enrich broadcasted content [13]. In the context of real-time automated production of team sports events [1,[6][7][8]11], knowing the ball position with accuracy and without delay is even more critical.…”