Hanwang Technology has obtained patents related to text recognition methods, improved the feature mapping layer, reduced parameters, and enhanced model speed.

On March 28, news from the National Intellectual Property Administration indicates that Hanwang Technology Co., Ltd. has applied for a patent titled “Multi-category Text Sequence Recognition Method, Device, and Electronic Equipment,” with the authorization announcement number CN116740730B, and the authorization announcement date set for March 27, 2026. The application publication number is CN116740730A, the application number is CN202310233711.2, the application publication date is March 27, 2026, the application date is March 3, 2023, and the inventors are Huang Wei and Liu Zhengzhen. The patent agency is Beijing Runzeheng Intellectual Property Agency Co., Ltd., with patent agent Ren Yajuan. The classification numbers are G06V30/18, G06V10/82, G06V30/19, G06N3/048, G06N3/0464, G06N3/08.

The patent abstract indicates that this application discloses a multi-category text sequence recognition method and device, belonging to the field of text recognition technology. The method is executed by a text sequence recognition model that includes at least one feature extraction layer, feature mapping layer, and transcription layer. The feature mapping layer comprises M parallel convolutional layers. The method includes: extracting features from the input image through the feature extraction layer to obtain N feature maps; distributing the N feature maps to the M convolutional layers for feature mapping, resulting in grouped feature mapping results; merging the grouped feature mapping results to obtain feature mapping results corresponding to a predetermined number of character categories; and transcribing the feature mapping results to obtain text sequence recognition results. This method improves the feature mapping layer by transforming it into multiple convolutional layers, grouping the feature maps outputted by the feature extraction layer for feature mapping, and then merging the feature mapping results, which reduces the number of parameters in the text sequence recognition model and enhances the model’s training and inference speed.

Hanwang Technology was established on September 11, 1998, and was listed on the Shenzhen Stock Exchange on March 3, 2010, with both its registered and office addresses in Beijing. The company is a leading enterprise in the domestic pattern recognition field, focusing its core business on intelligent interaction hardware and software, possessing several independent core technologies, with high technical barriers and significant investment value.

Hanwang Technology engages in the production of hardware and software products related to intelligent interaction centered around pattern recognition, industry applications, technology licensing, and platform services. The company falls under the Shenwan industry classification of computers - software development - vertical application software, involving sectors such as biometric recognition, online office, and voice technology.

In 2025, Hanwang Technology achieved an operating income of 1.889 billion yuan, ranking seventh among 14 companies in the industry, far below the first-place Desay SV’s 32.557 billion yuan and second-place Yunsai Zhihui’s 6.348 billion yuan, with the industry average being 4.867 billion yuan and the median at 1.849 billion yuan. Among its main businesses, the income from smart interactive pens was 1.129 billion yuan, accounting for 59.77%; AI terminal income was 500 million yuan, accounting for 26.47%; and multimodal big data business income was 230 million yuan, accounting for 12.15%. However, the company’s net profit was -188 million yuan, ranking last in the industry, with a noticeable gap compared to the first-place Tonghuashun’s 3.205 billion yuan and second-place Desay SV’s 2.473 billion yuan, with the industry average being 585 million yuan and the median at 161 million yuan.

The recent patent situation of Hanwang Technology Co., Ltd. is as follows:

Disclaimer: The market has risks, and investment requires caution. This article is automatically published by an AI large model based on third-party databases and does not represent the views of Sina Finance. Any information presented in this article is for reference only and does not constitute personal investment advice. Please refer to the actual announcements for discrepancies. For any questions, please contact biz@staff.sina.com.cn.

Massive information and precise interpretation, all available on the Sina Finance APP

Editor: Xiao Lang Kuai Bao