A novel tensor-information bottleneck method for multi-input single-output applications

Ensuring timeliness and mobility for multimedia computing is a crucial task wireless communication . Previous algorithms that utilize information channels, such as the bottleneck method, have shown great performance efficiency, which guarantees timeliness. However, methods suit only in handling single variable tasks image processing , but are inapplicable to multivariable applications video processing. To address this critical shortcoming, we propose novel tensor channel extends current single-input single-output matrix more practical multi-input channel. In comparison with classic channel, our not performs better experiments, also allows wider range of applications. We further build an innovative tensor-information method upon state-of-the-art method. Experiments on shot boundary detection conducted using benchmark data sets demonstrate effectiveness proposed approach compared methods. specific, yields 6.2% increase channel-based when other methods, achieve 0.1%–17.7% gains under different experimental configurations.