Temporally Identity-Aware SSD with Attentional LSTM

Temporal object detection has attracted significant attention, but most popular detection methods can not leverage the rich temporal information in videos. Very recently, many different algorithms have been developed for video detection task, but real-time online approaches are frequently deficient. In this paper, based on attention mechanism and convolutional long short-term memory (ConvLSTM), we propose a temporal signal-shot detector (TSSD) for real-world detection. Distinct from previous methods, we take aim at temporally integrating pyramidal feature hierarchy using ConvLSTM, and design a novel structure including a low-level temporal unit as well as a high-level one (HL-TU) for multi-scale feature maps. Moreover, we develop a creative temporal analysis unit, namely, attentional ConvLSTM (AC-LSTM), in which a temporal attention module is specially tailored for background suppression and scale suppression while a ConvLSTM integrates attention-aware features through time. An association loss is designed for temporal coherence. Besides, online tubelet analysis (OTA) is exploited for identification. Finally, our method is evaluated on ImageNet VID dataset and 2DMOT15 dataset. Extensive comparisons on the detection and tracking capability validate the superiority of the proposed approach. Consequently, the developed TSSD-OTA is fairly faster and achieves an overall competitive performance in terms of detection and tracking. The source code will be made available.