From October 23 to 24, the "2021 Artificial Intelligence the Taihu Lake Forum" hosted by the Jiangsu Artificial Intelligence Society was held in Wuxi. Xiaoshi was invited to participate, exchanging the latest research results with scholars and entrepreneurs in machine learning, artificial intelligence and related fields, conducting industry university research docking, sharing research results, innovative ideas, and the latest developments in the field of artificial intelligence.

The theme of this conference is "Frontier Development of New Generation Artificial Intelligence", aiming to gather experts and scholars in the field of artificial intelligence, conduct in-depth exchanges and discussions around the current hot topics, core technologies, and scientific issues of common concern in the field of artificial intelligence, and provide a platform for the exchange of the latest research results for the artificial intelligence industry in our province and across the country.

Xiaoshi CTO Hu Jianguo delivered a keynote speech on "AI Algorithm Scenario Implementation" at the forum, exploring how AI technology can better be implemented in various industries under the new infrastructure trend.

With artificial intelligence becoming a new trend, it is highly anticipated how various industries will collide with AI. To truly achieve the implementation of AI, it is necessary to delve into the industry field and organically integrate scenario based, industry data mechanisms, and AI algorithms.

小视专注智慧视觉领域6年，以深度学习算法，大数据分析、物联网平台为核心能力，为智慧城市、智慧社区、智慧校园、智慧工地、智慧交通、互联网等细分场景提供全生态服务。公司在全国各地已成功打造十余种场景的标杆案例，服务上千家客户，拥有上百家生态合作伙伴。

Minivision has been focusing on the field of smart vision for 6 years, with deep learning algorithms, big data analysis, and IoT platforms as its core capabilities, providing comprehensive ecological services for segmented scenarios such as smart cities, smart communities, smart campuses, smart construction sites, smart transportation, and the Internet. The company has successfully created benchmark cases for over ten scenarios across the country, serving thousands of customers, and has over a hundred ecological partners.

Hu Jianguo stated in the sharing that the implementation of AI algorithm scenarios cannot be achieved without computing power, algorithms, and data. Minivision is based on a small sample learning algorithm, which minimizes the amount of training data collected for long tail scenes as much as possible. At present, our self-developed algorithms cover various algorithms such as facial recognition, live body detection, human attributes, behavior analysis, machine object analysis, high-altitude throwing, industrial quality inspection, and portrait cartooning, to meet the needs of various scenarios.

The algorithm adapts to mainstream algorithm modules and board core solutions in the market, while supporting multiple systems such as Android, Windows, Linux, and localization. The advantage of Minivision lies in its terminal side support for lightweight deployment and excellent recognition accuracy and speed, as well as its performance tuning combined with powerful computing power and data to achieve rich deployment forms and flexible and convenient applications.

Taking the AI algorithm landing in the community scene as an example, Minivision's "Smart Community Full Stack Algorithm" aggregates more than ten popular algorithms, including high-altitude parabolic detection algorithm, dog walking without rope detection, non motorized vehicle entering elevator detection, business occupancy detection, face/human detection, and fall detection. The high-altitude parabolic detection algorithm solves multiple parabolic detection problems such as fast falling targets, small targets, and multiple targets, and can reduce interference factors such as dragonflies, birds, rain, and catkins.

The Smart Community Full Stack Algorithm combines intelligent cameras, small vision astrology edge analysis hosts, and real-time intelligent video analysis platforms to achieve one screen perception of the overall situation, comprehensively safeguard community security, truly define the eye of the community, and improve the level of community management services while making services more warm.

Minivision is committed to making AI more grounded and closely related to everyone, like the spring breeze moistening things, reaching out to thousands of industries within reach.