F-DES: Fast and Deep Event Summarization

Kumar, Krishan; Shrimankar, Deepti D.

doi:10.1109/tmm.2017.2741423

Cited by 128 publications

(30 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Jain et al propose a method that embeds structure into a deep model [31] to incorporate knowledge with deep models for activity recognition. Other deep approaches proposed for activity recognition are [32], [33]. A group of research has been conducted to use objects, their affordances, and states in video for action recognition [34], [35], [36], [37], [38], [39].…”

Section: B Video Understandingmentioning

confidence: 99%

“…Various researchers have proposed methods for handling multi-camera scenarios. Event summarization in multi-view videos using a deep learning approach [40], detection and summarization of an event in multi-view surveillance videos by applying boosting [41], and a machine learning ensemble method [42] are instances of research in the area of multi-view video understanding. We have not addressed this aspect of video understanding in this work.…”

Section: B Video Understandingmentioning

confidence: 99%

See 1 more Smart Citation

Long Activity Video Understanding Using Functional Object-Oriented Network

Jelodar

Paulius

Sun

2019

IEEE Trans. Multimedia

View full text Add to dashboard Cite

Video understanding is one of the most challenging topics in computer vision. In this paper, a four-stage video understanding pipeline is presented to simultaneously recognize all atomic actions and the single on-going activity in a video. This pipeline uses objects and motions from the video and a graphbased knowledge representation network as prior reference. Two deep networks are trained to identify objects and motions in each video sequence associated with an action. Low Level image features are then used to identify objects of interest in that video sequence. Confidence scores are assigned to objects of interest based on their involvement in the action and to motion classes based on results from a deep neural network that classifies the on-going action in video into motion classes. Confidence scores are computed for each candidate functional unit associated with an action using a knowledge representation network, object confidences, and motion confidences. Each action is therefore associated with a functional unit and the sequence of actions is further evaluated to identify the single on-going activity in the video. The knowledge representation used in the pipeline is called the functional object-oriented network which is a graph-based network useful for encoding knowledge about manipulation tasks. Experiments are performed on a dataset of cooking videos to test the proposed algorithm with action inference and activity classification. Experiments show that using functional object oriented network improves video understanding significantly.

show abstract

Section: B Video Understandingmentioning

confidence: 99%

Section: B Video Understandingmentioning

confidence: 99%

Long Activity Video Understanding Using Functional Object-Oriented Network

Jelodar

Paulius

Sun

2019

IEEE Trans. Multimedia

View full text Add to dashboard Cite

show abstract

“…Krishan Kumar et al [1] proposed FASTA approach which is local alignment based method to summarize the events in Multiview videos. Convolutional Neural Network (CNN) is trained with RGB input images with multiple multi-channel filters .Initially N frames of equal length of a single view is fed into these CNN to extract the visual features and object detection .…”

Section: J Fastamentioning

confidence: 99%

“…[8] outperforms well with [11] ¤ Rameswar Panda et al [3] proposed unsupervised framework via joint embedding and sparse representative selection to resolve inter-view dependencies among multi-view videos. They extract CNN visual features from the video frames .It outperforms well with the [11] and [8] ¤ Krishan Kumar et al [1] method extract the CNN features from the input video frame and nucleotide sequence is created based on the cosine similarity of CNN feature between the frames. FASTA algorithm is used to remove the inter-view redundancy and the correlations among multiple views is cptured using an optimized alignment approach.…”

Section: Inference From the Surveymentioning

confidence: 99%

See 1 more Smart Citation

Research on Damping Derivatives for Delta Wings in Hypersonic Flow for Curved Leading Edges with Full Sine Wave

Monis,

Crasta,

Faheem

et al. 2019

IJEAT

View full text Add to dashboard Cite

Video Surveillance System uses video cameras to capture images and videos that can be compressed, stored and send to place with the limited set of monitors .Now a Days all the public places such as bank, educational institutions, Offices, Hospitals are equipped with multiple surveillance cameras having overlapping field of view for security and environment monitoring purposes. A Video Summarization is a technique to generate the summary of entire Video Content either by still images or through video skim. The summarized video length should be less than the original video length and it should covers maximum information from the original video. Video summarization studies concentrating on monocular videos cannot be applied directly to multiple-view videos due to redundancy in multiple views. Generating Summary for Surveillance videos is more challenging because, videos Captured by surveillance cameras is long, contains uninteresting events, same scene recorded in different views leading to inter-view dependencies and variation in illuminations. In this paper, we present a survey on the research work carried on video summarization techniques for videos captured through multiple views. The summarized video generated can be used for the analysis of post-accident scenarios, identifying suspicious events, theft in public which supports Crime department for the investigation purposes.

show abstract

Face Mask Detection in Real‐Time Video Stream Using Deep Learning

Negi

Kumar

2021

Computational Intelligence and Healthcare Informatics

View full text Add to dashboard Cite

F-DES: Fast and Deep Event Summarization

Cited by 128 publications

References 39 publications

Long Activity Video Understanding Using Functional Object-Oriented Network

Long Activity Video Understanding Using Functional Object-Oriented Network

Research on Damping Derivatives for Delta Wings in Hypersonic Flow for Curved Leading Edges with Full Sine Wave

Face Mask Detection in Real‐Time Video Stream Using Deep Learning

Contact Info

Product

Resources

About