A visually disabled person-oriented automatic picture description method for web content barrier-free access

A technology for the visually impaired and web page content, applied in the field of automatic picture description, can solve problems such as difficulties, narrow bandwidth of screen reading software, lack of visual information assistance, etc., and achieve the effect of improving understanding

Inactive Publication Date: 2012-09-12

ZHEJIANG UNIV

0 Cites 10 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0004] 1) Screen reading software has narrow bandwidth and can only support sequential information access;

[0005] 2) Lack of visual information assistance, it is difficult to understand the content of the picture itself only from the picture instead of text description;

[0006] 3) Many webpage images lack alternative text, and the existing alternative text descriptions often lack useful information

[000...

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

Method used

4.1) word segmentation is carried out to the text of picture context, improve the weight o...

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

Abstract

The invention discloses a visually disabled person-oriented automatic picture description method for web content barrier-free access. The method comprises the steps of: grabbing tagged pictures from the network to form a sample library of annotated pictures; sorting the picture tags of the annotated picture sample library to realize tag denoising; generating tag recommendation words for illustrations of the webpages by using similar pictures of the sample library; and extracting sentences from contexts of the positions where the pictures are located on the WebPages in combination with tag words and using the sentences as descriptions of the pictures. The method comprehensively uses the tag ordering, tag recommending, and webpage document abstracting technology to generate brief descriptions for the pictures, and thus improves the webpage content barrier-free access for vision disabilities, thereby improving their web experience.

Application Domain

Special data processing applications

Technology Topic

Web contentFree access +4

Image

Examples

Experimental program(1)

Example Embodiment

[0033] The present invention will be further described below in conjunction with the drawings and embodiments.

[0034] 1. Grab tagged pictures on the Internet as a sample library of tagged pictures: Grab pictures and their corresponding tags from many current picture sharing websites. Commonly used picture sharing sources include: (1) Flickr, currently the most widely used picture sharing website, and provides a complete download API; you can use the Flickr API to grab a large number of pictures and corresponding tags. (2) LabelMe, this picture source provides a large number of high-quality label pictures.

[0035] 2. Sort the image tags in the annotated image sample library to achieve tag denoising:

[0036] 2.1) Use image color, texture, and shape features to measure the similarity between images, and use neighbor voting to sort image tags as the initial result of the sorting;

[0037] 2.2) Construct the jump probability matrix in the graph sorting algorithm according to the relationship between users, tags and pictures;

[0038] 2.3) Take the initial result of label sorting as the label preference in the graph sorting process, and use the graph sorting algorithm to refine the initial sorting result.

[0039] 3. Use the similar pictures in the sample library to generate label recommendations for the illustrations in the text: use color, texture and shape features to select pictures similar to the text illustration features from the sample library, and use the tags with the top similar images as the recommended tags Select the most frequent labels from the candidate words as the recommended labels for text illustrations.

[0040] 4. Combining tag words to extract sentences in the context of the text position of the picture as the description of the picture:

[0041] 4.1) Perform word segmentation on the text of the picture context, increase the weight of nouns, and increase the weight of the tag recommendation words and their synonyms matching vocabulary;

[0042] 4.2) Calculate the weight of the sentence in the picture context, and determine the weight of the sentence according to three aspects:

[0043] a) The sum of the weights of sentence vocabulary, in order to avoid the sentence is too long, divided by the number of sentence vocabulary;

[0044] b) The position of the sentence, the closer to the position of the picture, the greater the sentence weight coefficient: In actual operation, increase the sentence coefficient of the last sentence of the previous paragraph and the first sentence of the next paragraph at the text position of the picture;

[0045] c) The clue words of the sentence. If the sentence contains clue words related to the picture, the greater the weight coefficient of the sentence: In actual operation, increase the inclusion of "upper picture", "lower picture", "picture", "picture is The sentence weight of the clue word.

[0046] 4.3) The final picture description should be concise and contain as complete as possible relevant subject information. The final description consists of 2-3 sentences with the highest weight. In general, to ensure the conciseness of the picture description, the number of words is controlled within 50 words.

[0047] These embodiments are shown and described for the purpose of illustrating the function and structural principle of the present invention, and they can be modified without departing from such principle. Therefore, the present invention includes all modifications encompassed within the spirit and scope of the appended 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

PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

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

Classification and recommendation of technical efficacy words

Improve understanding

A visually disabled person-oriented automatic picture description method for web content barrier-free access

A technology for the visually impaired and web page content, applied in the field of automatic picture description, can solve problems such as difficulties, narrow bandwidth of screen reading software, lack of visual information assistance, etc., and achieve the effect of improving understanding

AI-Extracted Technical Summary

Problems solved by technology

Method used

Abstract

Image

Examples

Example Embodiment

PUM

Description & Claims & Application Information

Similar technology patents

Multiple-view display system with user recognition and operation method thereof

Divergent expanding type language study system and method

Spherical occulter coronagraph cubesat

Integrated maintenance scoring apparatus and method

Interactive outpatient quantity prediction visual analysis method and system based on CatBoost model

Classification and recommendation of technical efficacy words

Article of footwear having an upper with a structured intermediate layer

Fluid-filled baldder with a reinforcing structure

Systems and methods for remote computer-based analysis of user-provided chemogenomic data

Methods and apparatus for learning sensor data patterns for gesture-based input

Glove with stitchless seams