Diverse Sequential Subset Selection for Supervised Video Summarization

By Boqing Gong, Wei-Lun Chao, Kristen Grauman, Fei Sha. Neural Information Processing Systems, 2014

477 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/25da1b119ba1e0bb602be6ce8492d1e33dbac9ff

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Viewpoint-Aware Video Summarization (2018, background): “However, owing to the subjective property of this task, a supervised approach [21, 38, 32, 23, 12, 31, 13, 19, 9, 42], that trains a summarization model which takes human-created summaries as the supervision, became standard because of its better performance.”
Deep Attentive Video Summarization With Distribution Consistency Learning (2020, methodology): “Most approaches [2], [3], [6], [10] employ the frame-level importance scores as the ground truth, and thus, we follow this setting.”
Unsupervised Video Summarization with Adversarial LSTM Networks (2017, methodology): “We believe that this is mainly because of the fact that unlike the repelling regularizer, DPP is non-linear and can reinforce stronger regularization.”
Video Summarization Using Deep Neural Networks: A Survey (2021, methodology): “This protocol was also used in the supervised video summarization approach of [96], while a variation of it was used in [97], [98], [99].”
Learning to Diversify Deep Belief Networks for Hyperspectral Image Classification (2017, background): “The typical applications of the diversified models include image restoration [36], image segmentation [34], image classification [34], [38], and video summarization [35].”
A Practical Online Framework for Extracting Running Video Summaries under a Fixed Memory Budget (2021, methodology): “In the literature, static summarization has often been approached in the oﬄine setting [7, 18, 10].”
Video Summarization with Long Short-Term Memory (2016, methodology): “Alternatively, the determinatal point process (DPP) — a probabilistic model that characterizes how a representative and diverse subset can be sampled from a ground set — is a valuable tool to model summarization in the supervised setting [2,16,18].”
Fair and Diverse DPP-based Data Summarization (2018, methodology): “DPP-based sampling has been deployed for many data summarization tasks including text and images (Kulesza & Taskar, 2011), videos (Gong et al., 2014), documents (Lin & Bilmes, 2012), recommendation systems (Zhou et al.”
Efficient Video Classification Using Fewer Frames (2019, background): “This has led to an increasing interest in automatic video processing with the aim of identifying activities [32, 40], generating textual descriptions [9], generating summaries [11, 27], answering questions [15] and so on.”
Enhancing Video Summarization via Vision-Language Embedding (2017, methodology): “(a) Baselines, including our reproduction of [7, 9] using their code with updated visual features.”
SCSampler: Sampling Salient Clips From Video for Efficient Action Recognition (2019, background): “For example, video summarization [18, 19, 29, 37, 56, 57, 58] and the automatic production of sport highlights [30, 31] entail creating a much shorter version of the original video by concatenating a small set of snippets corresponding to the most informative or exciting moments.”
Pano2Vid: Automatic Cinematography for Watching 360° Videos (2016, background): “A summary can take the form of a keyframe sequence [8–13], a sequence of video highlight clips [14–19], or montages of frames [20] or video clip excerpts [21].”
DLow: Diversifying Latent Flows for Diverse Human Motion Prediction (2020, background): “ons [7] and multiple choice learning [28,44]. Alternatively, submodular function maximization has been applied to select a diverse subset of garments from fashion images [31]. Another type of methods [40,20,19,32,5,74,70] seeks diversity using determinantal point processes (DPPs; [50,41]) which are ”
Textually Customized Video Summaries (2017, methodology): “Supervised methods use summary annotations obtained from human workers in the form of ground truth summaries [7, 8], highlight annotations [45] or GIF-formatted summaries [9].”
Diversity and Inclusion Metrics in Subset Selection (2020, background): “as biology [4, 18, 23] and ecology [21, 34, 35]. Geometric or distance-based measures of diversity have also been explored within the sciences, measuring the diversity of a datasetbythedataset’svolume[1,8,11,17,20,22,37],varianceasin PCA [30], or other measures of spread. The notion of heterogeneity”
Making 360° Video Watchable in 2D: Learning Videography for Click Free Viewing (2017, background): “Video summarization Video summarization aims to generate a concise representation for a video by removing temporal redundancy while preserving the important events [8, 14–16, 24, 27, 32, 33, 38, 42].”
Subset Replay Based Continual Learning for Scalable Improvement of Autonomous Systems (2018, background): “Choosing the right subset has been a critical function for other associated problems like document or video summarization [7] and active learning.”
Video Summarization With Attention-Based Encoder–Decoder Networks (2017, background): “For example, Gong et al. [14] treat video summarization as a supervised subset selection problem, and present a probabilistic model called sequential Determinatal Point Process (seqDPP) to learn how a diverse and representative subset is selected from the training set.”
Fast Greedy MAP Inference for Determinantal Point Process to Improve Recommendation Diversity (2017, methodology): “Despite these theoretical guarantees, greedy algorithm is widely used for DPP due to its promising empirical performance [29, 19, 49].”