High-Level Descriptors for Fall Event Detection Supported by a Multi-Stream Network

The need for assertive video classification has been increasingly in demand. Especially detecting endangering situations, it is crucial to have a quick response avoid triggering more serious problems. During this work, we target concerning falls. Our study focuses on the use of high-level descriptors able correctly characterize event. These descriptor results will serve as inputs multi-stream architecture VGG-16 networks. Therefore, our proposal based analysis best combination extracted features binary videos. This approach was tested three known datasets, and proven yield similar other consuming methods found literature.