Video super-resolution using an adaptive superpixel-guided auto-regressive model

Li, Kun; Zhu, Yanming; Yang, Jingyu; Jiang, Jianmin

doi:10.1016/j.patcog.2015.08.008

Cited by 52 publications

(17 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More recent research has also tried to produce high-quality saliency maps in a filtering-based framework [62]. Such salient object segmentation has achieved great success for simple images in image scene analysis [63][64][65], and content aware image editing [66,67]; it can be used as a cheap tool to process a large number of Internet images or build robust applications [68][69][70][71][72][73] by automatically selecting good results [61,74]. However, these approaches are less likely to work for complicated images in which many objects are present but are rarely dominant (e.g., PASCAL VOC images).…”

Section: Salient Object Detectionmentioning

confidence: 99%

BING: Binarized normed gradients for objectness estimation at 300fps

et al. 2019

View full text Add to dashboard Cite

Training a generic objectness measure to produce object proposals has recently become of significant interest. We observe that generic objects with well-defined closed boundaries can be detected by looking at the norm of gradients, with a suitable resizing of their corresponding image windows to a small fixed size. Based on this observation and computational reasons, we propose to resize the window to 8 × 8 and use the norm of the gradients as a simple 64D feature to describe it, for explicitly training a generic objectness measure. We further show how the binarized version of this feature, namely binarized normed gradients (BING), can be used for efficient objectness estimation, which requires only a few atomic operations (e.g., add, bitwise shift, etc.). To improve localization quality of the proposals while maintaining efficiency, we propose a novel fast segmentation method and demonstrate its effectiveness for improving BING's localization performance, when used in multithresholding straddling expansion (MTSE) postprocessing. On the challenging PASCAL VOC2007 dataset, using 1000 proposals per image and intersectionover-union threshold of 0.5, our proposal method achieves a 95.6% object detection rate and 78.6% mean average best overlap in less than 0.005 second per image.

show abstract

Section: Salient Object Detectionmentioning

confidence: 99%

BING: Binarized normed gradients for objectness estimation at 300fps

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The following frames eliminate the motion and are used as CNN inputs that offer excellent video output as output. Only Li et al [12] have developed a large resolution model based on pixel models. The keyframe is automatically selected and has a rare withdrawal method.…”

Section: E Resolution Enhancementmentioning

confidence: 99%

Video Enhancement Techniques Employed Over the Decade for Reliable Communication

Reddy¹,

Vinayakamurthy²,

Barath³

2019

IJITEE

View full text Add to dashboard Cite

Traditional networks are designed to meet the needs of streaming video on the Internet. Relatively high standards of broadband connections and video broadband services make the best effort for certain specific video information and transmission methods to appear at any time to reduce network constraints. Video quality is directly related to the basic ability of the phone to send and receive data from time to time for playback. Record videos from your mobile phone are facing various challenges that have been overcome by a variety of methods. In an extremely robust and unpredictable environment, transversal hardware provides adequate video. Conventional networks are designed to meet the needs of streaming video over the Internet. Relatively high standards of broadband connections and video broadband services make the best effort for certain specific video information and transmission methods to appear at any time to reduce network constraints. Video quality is directly related to the basic ability of the phone to send and receive data from time to time for playback. Record videos from your mobile phone are facing various challenges that have been overcome by a variety of methods. In an extremely robust and unpredictable environment, transversal hardware provides adequate video.

show abstract

“…Frames of a video sequence may contain many small shifted or rotated LR images of a given object, caused by the acquisition process, and the camera and/or scene motion, from which a HR image can be obtained using MISR techniques. MISR can be applied to obtain either a single HR image from a sequence of many LR images or a HR image sequence from a LR image sequence [42][43][44][45].…”

Section: State-of-the-artmentioning

confidence: 99%

Image super-resolution for outdoor digital forensics. Usability and legal aspects

Villena

Vega

Mateos

et al. 2018

Computers in Industry

View full text Add to dashboard Cite

Digital Forensics encompasses the recovery and investigation of data, images, and recordings found in digital devices in order to provide evidence in the court of law. This paper is devoted to the assessment of digital evidence which requires not only an understanding of the scientific technique that leads to improved quality of surveillance video recordings, but also of the legal principles behind it. Emphasis is given on the special treatment of image processing in terms of its handling and explanation that would be acceptable in a court of law. In this context, we propose a variational Bayesian approach to multiple-image super-resolution based on Super-Gaussian prior models that automatically enhances the quality of outdoor video recordings and estimates all the model parameters while preserving the authenticity, credibility and reliability of video data as digital evidence. The proposed methodology is validated both quantitatively and visually on synthetic videos generated from single images and real-life videos and applied to a real-life case of damages and stealing in a private property.

show abstract

Video super-resolution using an adaptive superpixel-guided auto-regressive model

Cited by 52 publications

References 50 publications

BING: Binarized normed gradients for objectness estimation at 300fps

BING: Binarized normed gradients for objectness estimation at 300fps

Video Enhancement Techniques Employed Over the Decade for Reliable Communication

Image super-resolution for outdoor digital forensics. Usability and legal aspects

Contact Info

Product

Resources

About