Unlock instant, AI-driven research and patent intelligence for your innovation.
Image identification method and device
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
An image recognition, image technology, applied in character and pattern recognition, instruments, computer parts and other directions, can solve the problem of unsatisfactory recognition effect, and achieve the effect of high accuracy and precision
Inactive Publication Date: 2013-04-10
GUANGDONG TUTUSOU NETWORK TECH
View PDF3 Cites 9 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0004] The above method is suitable for objects to be recognized with relatively simple features. When the shape of the object to be recognized is changeable, such as the human body, the recognition effect of the above method is not ideal
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
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0023] The image recognition method of this embodiment, such as figure 1 shown, including the following steps:
[0024] Step S101 , divide the image, extract the gradient histogram features of each small square after division, and the number of divisions is balanced between the recognition efficiency and recognition accuracy of the image recognition method. Assuming that each image is divided into m small squares, the m small squares are recorded as g i ,i=1,2...m.
[0025] Step S102, establishing a model based on shape estimation that is suitable for the object to be recognized, and the model is composed of various parts. For example, if the entire model consists of k parts, each part is denoted as p k ,k=1,2...K. The local scores corresponding to each part of the model are calculated for each small square according to the gradient histogram features.
[0026] The partial scoring formula is: S k (I, g i ,p k )=w k *φ(I, g i )+b k
[0027] Among them, I represents ...
Embodiment 2
[0039] In order to further improve the recognition results, this embodiment adds the judgment of the position of the object to be recognized on the basis of Embodiment 1. Specifically, after judging that there is an object to be recognized in the image, the middle and small squares are combined according to the small grid with the highest global score. The coordinates of the grid determine the position of the object to be recognized in the image. In this way, the combination of small squares at corresponding positions can be marked in the image as the object to be recognized for reference by the user.
[0040] Other technical features of this embodiment are the same as those of Embodiment 1, and will not be repeated here.
Embodiment 3
[0042] If the size of the image does not match the size of the model, it may cause false recognition consequences. Therefore, in this embodiment, the image can also be down-sampled, and the steps from S101 to S104 are performed on each image after the down-sampling, then each image has a highest global score, and then the highest global score and the highest global score are selected. The threshold value is compared and judged happily. Images of various sizes obtained after downsampling, that is, image pyramids, at least one size of the image matches the size of the model. Therefore, using this embodiment can further increase the accuracy and precision of image recognition.
[0043] For the down-sampling performed in this embodiment, the more stages and the finer the recognition effect, the more accurate the recognition effect, but the recognition efficiency is sacrificed. Therefore, the number of down-sampling stages also depends on requirements, preferably ten stages. In a...
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
Login to View More
Abstract
The invention discloses an image identification method and an image identification device. A model based on morphology estimates is established in advance. The model is adaptive to objects to identify and composed of each part. The image is divided into a plurality of small checks. At first the small checks which are matched with parts of the model are found according to local scores, then the small checks are assembled. Various kinds of possible objects to identify are acquired, and assemblies which are not in coincidence to spatial logic relationship are eliminated. An assembly with highest overall score is found from the rest assemblies. If the overall score of the assembly with highest overall score is over the threshold, that the object to identify exists in the image is judged. Due to the fact that the image identification method and the image identification device use the model based on the morphology estimates as reference, and the images is divided and reassembled. Whether the object to identify exists in the images or not can be ensured through two layer screens of the local score and the overall score. People or objects with complicated morphological characteristics can be identified, and at the same time the image identification method and the image identification device have higher accuracy and precision.
Description
technical field [0001] The invention relates to the technical field of digital imageprocessing, in particular to an image recognition method and device. Background technique [0002] Image recognition is a technology that recognizes images through semantic information in images, and its application fields are very broad, such as traffic sign recognition, license plate recognition, face recognition, and medical image recognition. [0003] Most of the traditional image recognition methods are based on the visual dictionary model. Such methods first need to extract the local features of the image, and then perform cluster analysis on the local features to establish a visual vocabulary. This allows an image to be represented as a histogram based on visual vocabulary. Finally, a machine learning method is used to train a recognizer for predicting images. [0004] The above method is suitable for objects to be recognized with relatively simple features. When the shape of the ob...
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
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More