VSIIR: VNIIRS Semantics Inference for Interpretability and Rating

The Video-National Imagery Interpretability Rating Scale (VNIIRS) is a task-based scale that reflects observable semantic content in videos. Intelligence agencies such as the NGA use VNIIRS for applications such as efficient storage and retrieval, data compression and data browsing. Unlike video quality metrics, VNIIRS ratings reflect subjective assessment of semantic content, which may not always be correlated with video quality metrics such as blur and noise. An automated, real-time tool that can annotate videos using its semantic content and estimate the VNIIRS rating can help analysts to analyze and search video data quickly and efficiently. \n\n The novel VSIIR framework uses state-of-the-art deep learning based techniques to estimate VNIIRS rating. The approach extracts robust temporal and semantic features from video to estimate the VNIIRS rating. The benefits of VNIIRS rating will be demonstrated using a new process framework for continuous learning and object tagging as an application.