No abstract

No abstract

Abstract. Object proposals have contributed significantly to recent advances in object understanding in images. Inspired by the success of this approach, we introduce Deep Action Proposals (DAPs), an effective and efficient algorithm for generating temporal action proposals from long videos. We show how to take advantage of the vast capacity of deep learning models and memory cells to retrieve from untrimmed videos temporal segments, which are likely to contain actions. A comprehensive evaluation indicates that our approach outperforms previous work on a large scale action benchmark, runs at 134 FPS making it practical for large-scale scenarios, and exhibits an appealing ability to generalize, i.e. to retrieve good quality temporal proposals of actions unseen in training.

In this work, we present a new intuitive, end-to-end approach for temporal action detection in untrimmed videos. We introduce our new architecture for Single-Stream Temporal Action Detection (SS-TAD), which effectively integrates joint action detection with its semantic sub-tasks in a single unifying end-to-end framework. We develop a method for training our deep recurrent architecture based on enforcing semantic constraints on intermediate modules that are gradually relaxed as learning progresses. We find that such a dynamic learning scheme enables SS-TAD to achieve higher overall detection performance, with fewer training epochs. By design, our single-pass network is very efficient and can operate at 701 frames per second, while simultaneously outperforming the state-of-the-art methods for temporal action detection on THUMOS'14.

In this paper, we introduce the task of retrieving relevant video moments from a large corpus of untrimmed, unsegmented videos given a natural language query. Our task poses unique challenges as a system must efficiently identify both the relevant videos and localize the relevant moments in the videos. This task is in contrast to prior work that localizes relevant moments in a single video or searches a large collection of already-segmented videos. For our task, we introduce Clip Alignment with Language (CAL), a model that aligns features for a natural language query to a sequence of short video clips that compose a candidate moment in a video. Our approach goes beyond prior work that aggregates video features over a candidate moment by allowing for finer clip alignment. Moreover, our approach is amenable to efficient indexing of the resulting clip-level representations, which makes it suitable for moment localization in large video collections. We evaluate our approach on three recently proposed datasets for temporal localization of moments in video with natural language extended to our video corpus moment retrieval setting: DiDeMo [16], Charades-STA [10], and ActivityNetcaptions [22]. We show that our CAL model outperforms the recently proposed Moment Context Network [16] on all criteria across all datasets on our proposed task, obtaining an 8%-85% and 11%-47% boost for average recall and median rank, respectively, and achieves 5× faster retrieval and 8× smaller index size with a 1M video corpus.