Company News

News Center

Bring together comprehensive, cutting-edge and in-depth information and media focused reports

The results of the battle have been announced and we have won a silver medal! Xiaoshi Team Competes in the 2022 Ascension AI Innovation Competition National Finals

Company News 2022-11-17 2232 views

Discover developers and ignite the power to change the world.

From November 15th to 16th, the 2022 Shengteng AI Innovation Competition National Finals were successfully held in the ancient capital of Xi'an. The "Minivision" team of Minivision Technology represented the Jiangsu division and won the "Silver Award" in the application track with the "Minivision Intelligent Vision Central Platform" solution.


照片 (4).jpg


This competition is jointly organized by more than 20 Shengteng Ecological Innovation Centers across the country, Huawei, the New Generation Artificial Intelligence Industry Technology Innovation Strategic Alliance, OpenI Qizhi Community, and the China Artificial Intelligence Industry Development Alliance.


照片 (2).jpg


The competition consists of two major tracks: application and promotion. The application track where the "cannot be underestimated" team is located aims to explore innovative solutions with specific implementation based on the Shengteng AI software and hardware platform, facing industry scenarios.


AI support accelerates the digitization of urban governance


The Minivision Intelligent Vision Central Platform is equipped with Ascension Inference Server and Ascension Training Resources at the basic software and hardware layer, which can fully empower city level and large number of video monitoring equipment, and assist in urban "one network management".


111.png


By providing an open interface, the platform enables quick linking of scene ecology. Based on the MindSpot framework, the output efficiency of the algorithm model has been improved by at least 6 times. By using the MindX series of tools, the model upgrade iteration efficiency has been improved by 50%, and the algorithm implementation is more efficient.

To enhance the adaptability and robustness of multiple scenarios, Minivision adopts self-developed small sample robust learning technology, heuristic model feature fusion technology, and cross domain simulation technology to improve algorithm accuracy in the case of limited training data.


222.png

Relying on the capabilities of high-capacity video detection and video structuring technology, multi object tracking and detection technology, and self attention mechanism technology, we can quickly and efficiently analyze and develop video structuring for algorithmic scenarios. In complex scenarios, the platform can achieve stable recall of small targets and robust tracking of multiple targets.


333.png



The Minivision Intelligent Vision Central Platform can be replicated on a large scale and has a variety of landing scenarios. It has already been deployed on a large scale in multiple provinces and cities such as Guangdong, Jiangsu, Anhui, Zhejiang, Shanghai, and Gansu, assisting in various aspects of local urban governance.


Focusing on the future of digital intelligence, achieving stability is the key to achieving success. Minivision Technology will always be committed to the research and development of innovative technologies in the "AI+industry", providing more forward-looking and innovative application products and solutions for the construction of smart cities, and expanding new space for the development of the artificial intelligence industry.