Advances in visual perceptual tasks have been mainly driven by the amount, and types, of annotations of large scale datasets. Researchers have focused on fully-supervised settings to train models using offline epoch based schemes. Despite the evident advancements, limitations and cost of manually annotated datasets have hindered further development for event perceptual tasks, such as detection and localization of objects and events in videos. The problem is more apparent in zoological applications due to the scarcity of annotations and length of videos - most videos are at most ten minutes long. Inspired by cognitive theories, we present a self-supervised perceptual prediction framework to tackle the problem of temporal event segmentation by building a stable representation of event-related objects. The approach is simple but effective. We rely on LSTM predictions of high-level features computed by a standard deep learning backbone. For spatial segmentation, the stable representation of the object is used by an attention mechanism to filter the input features before the prediction step. The self-learned attention maps effectively localize the object as a side effect of perceptual prediction. We demonstrate our approach on long videos from continuous wildlife video monitoring, spanning multiple days at 25 FPS. We aim to facilitate automated ethogramming by detecting and localizing events without the need for labels. Our approach is trained in an online manner on streaming input and requires only a single pass through the video, with no separate training set. Given the lack of long and realistic (includes real-world challenges) datasets, we introduce a new wildlife video dataset – nest monitoring of the Kagu (a flightless bird from New Caledonia) – to benchmark our approach. Our dataset features a video from 10 days (over 23 million frames) of continuous monitoring of the Kagu in its natural habitat. We annotate every frame with bounding boxes and event labels. Additionally, each frame is annotated with time-of-day and illumination conditions. We will make the dataset, which is the first of its kind, and the code available to the research community. We find that the approach significantly outperforms other self-supervised, traditional (e.g., Optical Flow, Background Subtraction) and NN-based (e.g., PA-DPC, DINO), baselines and performs on par with supervised boundary detection approaches (i.e., PC). At a recall rate of 80%, our best performing model detects one false positive activity every 50 minutes of training. On average, we at least double the performance of self-supervised approaches for spatial segmentation. Additionally, we show that our approach is robust to various environmental conditions (e.g., moving shadows). We also benchmark the framework on other datasets (i.e., Kinetics-GEBD, TAPOS) from different domains to demonstrate its generalizability.
Overview of our full architecture. The perceptual processing unit encodes current frames and future frames into a grid feature representations. An attention operation is applied to the current features to spatially segment the event objects. The predictor combines the event model representation with the current features to predict the future features. Error in the prediction is used as a learning signal for the trainable weights. The spatio-temporal pooling layer receives as input spatial localization map and prediction error signal and outputs the detected events.
This research is supported by the US National Science Foundation grants CNS 1513126 and IIS 1956050. The bird video dataset used in this paper was made possible through funding from the Polish National Science Centre (grant NCN 2011/01/M/NZ8/03344 and 2018/29/B/NZ8/02312). Province Sud (New Caledonia) issued all permits required for data collection.
© This webpage was in part inspired from this template.