Scene video text tracking method based on energy minimization

An energy minimization, text technology, applied in the field of computer vision, can solve the problems of blurred jitter, dense target, large scale change, etc.

Inactive Publication Date: 2019-04-05
UNIV OF SCI & TECH BEIJING
View PDF4 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The text in the scene video has the characteristics of large scale changes, dense objects, affected by lighting ...

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Scene video text tracking method based on energy minimization
  • Scene video text tracking method based on energy minimization
  • Scene video text tracking method based on energy minimization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0126] (1) datasets used in the present invention

[0127] The present invention first conducts experiments on the Minetto data set. The Minetto dataset was released when Minetto introduced its scene text detection and recognition method in Document 3. The dataset contains 5 video clips: Bateau, Bistor, Cambronne, Navette and Zara, all of which are common scenes in daily life. These The technical requirements for fragment inspection are different: Bateau examines the tracking situation when text is moved in translation, Bistor examines the influence of light and shadow, Cambronne examines the influence of many targets at the same time, Navette examines the tracking situation when the size of the target changes, and Zara examines strong light Impact on trackers.

[0128] The experiment of the present invention is also carried out on the ICDAR13-CH3-TASK1 data set. The ICDAR Robust Reading Competition included video text positioning and tracking as a new challenge in the compe...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a scene video text tracking method based on energy minimization, and relates to the field of computer vision. The method comprises the steps of describing a linear energy function of a fusion detection energy model, an apparent energy model and a mutual exclusion energy model of a tracking model; Solving the tracking model; performing Specific tracking process. The invention proposes a new text tracking method for scene video, and designs a corresponding detection energy model, an apparent energy model and a mutual exclusion energy model for the text of the scene video.By integrating the above three, overall energy is obtained. The optimal tracking trajectory set is obtained by minimizing the overall energy. It is considered that the trajectory set has the smallestloss energy and is overall optimal.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a method for tracking scene video text based on energy minimization. Background technique [0002] The semantic information contained in the text in the video plays an important role in video analysis and retrieval. It can be applied in many aspects, such as real-time translation system, assisted driving system and robot navigation system, etc. In recent years, text tracking of scene videos has also attracted the attention of scholars in related fields, and some research results have been achieved in this area. Most of the current text tracking methods belong to the Tracking-by-detection framework. Tracking-by-detection can be regarded as a data association problem, that is, the detection information of adjacent frames is associated together through tracking. Document 1 (Tanaka M, Goto H. Autonomous Text Capturing Robot Using Improved DCT Feature and TextTracking[C] / / Internationa...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00
CPCG06V20/41G06V30/10
Inventor 殷绪成蒙丽宇裴唯一田澍朱超
Owner UNIV OF SCI & TECH BEIJING
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap