The role of OCR technology in witness terminals OCR generally includes two steps: 1. Detection-->Find the area that contains the text (Proposal); 2. Classification-->Identify the text in the area.
Text Detection algorithms have developed rapidly in recent years and can be divided into horizontal text detection (Four degrees of freedom, similar to object detection)And tilt text detection (
The text frame is an irregular quadrilateral with eight degrees of freedom)
Text recognition has not made much progress in the past two years. * The direction of recent popularity is to put text detection and recognition into a network. First, OCR is an area of pattern recognition, so the whole process is also the process of pattern recognition.
As a whole, the process can be divided into the following steps: preprocessing: processing images containing characters for subsequent feature extraction and learning.
The main purpose of this process is to reduce the feature extraction and dimensionality reduction of useless information in images: features are the key information used to identify characters, and each different character can be distinguished from other characters through features.
For numbers and English letters, this feature extraction is relatively easy.
Classifier design, training and actual recognition: the classifier is used for recognition, that is, for the second step, you extract features from a text image and throw them to the classifier, the classifier classifies it and tells you which text this feature should recognize.
Post-processing: Post-processing is used to optimize the classification results, *, the classification of the classifier is sometimes not necessarily completely correct (In fact, it is not completely correct).
Second, OCR recognition images often have a large number of characters. In post-processing, you can try to format the recognition results, such as arranging according to the layout of the images. For example: this is a high-speed camera that uses OCR technology to text the identity line.
Shenzhen ptksai Technology Co. , Ltd. was founded in 2009. It is a fast-growing high-tech enterprise group. Its products mainly include: intelligent touch Android all-in-one machine, intelligent touch POS machine, touch Windows intelligent terminal, intelligent digital photo frame, stand-alone and online advertising machine, industrial network equipment and commercial display equipment, video greeting card, professional clock machine, and customized electronic products for major customers.
The company has three technology centers combining production, teaching and research in Xi 'an, Shenzhen and Silicon Valley of the United States. 61 engineers are engaged in software and hardware customization and development across time zones, as well as product front-end research.
In 2017, the company's R & D team contributed nine innovative invention patents.
The high-efficiency intelligent refrigerator and touch POS terminal developed by the company are at the advanced level of the industry. The company is also a solution company for a number of intelligent advertising machines and intelligent digital photo frames, providing complete machines, boards and overall solutions.
The company has three manufacturing bases in Shenzhen Longhua, Shenzhen Shajing and Dongguan Daling Mountain, covering an area of 21500 m² square meters. It has its own high-speed Mounter, injection molding machine, mold workshop and dust-free automatic production line. The annual production capacity of PCBA board is 2. 3 million pieces, the annual output of the intelligent terminal machine is 1. 1 million units, which can provide OEM/ODM/EMS services for world-renowned brands.
At present, the company has passed the certification of quality assurance system ISO9001 and environmental certification system ISO14001, and its products have passed CE, FCC, CCC, U/L, EUP, Rohs, Reach and other certifications. The company looks forward to long-term cooperation with friends at home and abroad. .