上海娱乐导航 – 第 2 页 – 上海后花园论坛,上海419论坛,爱上海龙凤419

China smart car research selected for CVPR2023! A new method for 3D lane line detection, produced by Mimo Zhixing.

2023年4月16日 | admin | restert

CVPR 2023, one of the top three computer vision conferences, was exposed in the paper of China Smart Car Company.

Domestic autonomous driving companyHao mo zhi hangThe paper Bev-Lane Det: A Simple and Effective 3D Lane Detection Baseline submitted by the technical team of its Artificial Intelligence Center was successfully selected into CVPR2023.

What problems should the new research solve?

3D lane detectionAs a hot topic in the field of automatic driving, it plays a vital role in vehicle routing. In the past, related technologies were often difficult to be applied because of its complex spatial transformation and inflexible representation of 3D lanes.

Faced with these problems, Millie independently developed an efficient and robust monocular 3D lane detection, referred to as:BEV-LaneDet.

In BEV-LaneDet, Millimeter team first converted all the internal and external parameters of the input image into unified internal and external parameters through the virtual camera, which ensured the consistency of the spatial relationship of the front cameras in different vehicles and effectively promoted the learning process.

Moreover, Millimeter Team uses feature extractor to extract the features of the front view image, and proposes a simple but effective 3D lane representation, which is called "Key-Points Representation". This module is more suitable for representing complex and diverse 3D lane structures.

On this basis, inspired by FPN, Millicent team also designed a lightweight and chip-friendly spatial transformation module called "Spatial Transformation Pyramid", which is a fast multi-scale spatial transformation module based on VPN, mainly responsible for the transformation from front view features to BEV features.

Finally, the team predicted the lane on the plane tangent to the local road surface, divided the local road surface into s1×s2 grids, predicted confidence, embedding for clustering, lateral error offset and the height of each grid, and used fast clustering method to fuse the results of each branch in reasoning to obtain 3D lanes.

How did BEV-LaneDet perform?

Through the above methods, the F1-S?core of BEV-LaneDet can reach 185FPS on the OpenLane dataset, which is 10.6% higher than the current SOTA.

BEV-LaneDet can also lead the current SOTA method by 4.0% in the performance of Apollo 3D Lane Synthetic.

At present, Mimo Zhixing is the first echelon player in domestic self-driving mass production, and the paper written by its technical team BEV-LaneDet was selected for CVPR, which is actually the embodiment of the technical strength behind the commercial progress.

At present, HPilot has iterated three generations of products, and has achieved mass production on nearly 20 models of Wei brand, tank, Haval, Euler, Great Wall Gun and other brands. In the future, passenger cars equipped with Millimeter-assisted driving products will reach millions of orders of magnitude. As of February 2023, the mileage of user-assisted driving has exceeded 35 million kilometers. NOH, the first city in China that can land on a large scale, has also reached the delivery status with its software sealed.

In addition, the terminal logistics automatic delivery vehicle has initially completed the commercial closed loop, delivering more than 1,000 units, and the delivery order volume of the terminal logistics automatic delivery vehicle Little Magic Camel series has also exceeded 160,000, and the commercialization process is accelerating in an all-round way.

Introduction to CVPR

IEEE (Institute of Electrical and Electronics Engineers) CVPR (Conference on Computer Vision and Pattern Recognition) is one of the three top conferences in the field of computer vision and pattern recognition.

The main content of the conference is computer vision and pattern recognition technology. CVPR is the world’s top computer vision conference (one of the three top conferences, the other two are ICCV and ECCV). In recent years, there are about 1,500 participants every year, and the number of papers collected is generally about 300.

CVPR has strict employment standards. According to CVPR officials, CVPR2023 received a total of 9,155 papers, a record high, and the organizing committee finally hired 2,360 papers, with an acceptance rate of 25.78%.

CVPR is also the witness of major changes in the field of AI vision. The current wave of AI driven by deep learning originated from Hinton’s blockbuster in CVPR Challenge.

Address: https://arxiv.org/pdf/2210.06006.pdf.

— finish —

[Smart Car Reference] Original content is prohibited from being reproduced at will without the authorization of the account.

Awesome! Shenzhen University pioneered：＂Virtual Digital Man＂ micro-specialty!

2023年4月9日 | admin | restert

Development of 5G, AI, VR and other technologies

The "Virtual Digital Man" industry is booming.

What? Virtual digital people?

I’ve heard of virtual numbers, I’ve heard of people,

Why can’t you understand this when you combine it?

Don’t worry!

This is not a word mashup, but the first micro-specialty in Shenzhen University.

-"Virtual Digital Man"

With the advent of the "Meta-Universe" era, virtual human technology has gradually become one of the hottest industrial tracks.Recently, Shenzhen Datong has approved seven micro-professional construction projects, including "Virtual Digital Man", "Edge Computing and Internet of Things Communication".Among them, the micro-specialty of "Virtual Digital Man" is a school-enterprise cooperation between Communication College and Tencent Technology (Shenzhen) Co., Ltd., which is committed to breaking down the barriers between colleges and disciplines and cooperating to cultivate compound innovative talents in short supply in the market.

"Virtual Digital Man"

"Virtual digital human" is a visible, interactive and adjustable virtual human form that digitizes the human body structure and presents it on the terminal screen through computer technology. As a hot field at present, it is the intersection of computer, digital media, marketing and other disciplines, and it is also the new direction of head Internet companies such as Tencent, Baidu and Iflytek.

From a technical point of view, virtual digital people can be divided into two categories: human-driven and intelligent-driven. Reality-driven is a relatively mature field at present, represented by Tianyi Luo, Liu Yexi, Xing Pupil and AYAYI in the industry. At present, the related concepts of "Metauniverse" form a clustering effect, and virtual digital human is one of the core industrial links of Metauniverse. With the lowering of the threshold and the great richness of application scenarios, it is estimated that by 2025, the virtual digital human industry will reach 100 billion.

In March, 2021, the state incorporated the development of virtual digital technology into the 14th Five-Year Plan for National Economic and Social Development of People’s Republic of China (PRC) and the Outline of Long-term Goals in 2035. Realizing virtual digital technology innovation has become the only way for China to realize industrial innovation and technological power in the future. In this context, Shenzhen University grasped the reality and focused on the future, and took the lead in creating a "virtual digital person" micro-specialty.

Based on the school-enterprise cooperation between the School of Communication of Shenzhen University and Tencent Technology (Shenzhen) Co., Ltd., the micro-specialty of "Virtual Digital Man" aims to meet the audio-visual content needs of the meta-universe and artificial intelligence, and cultivate the production methods and principles of virtual digital man. Master the planning, creation and communication skills of virtual IP; Know some artificial intelligence, algorithms and knowledge mapping principles, and be able to integrate code language with images; Establish professional innovative talents with professional understanding and research ability in the meta-universe industry, so as to get through the "last mile" of professional education and professional needs.

Micro-courses will adopt the training mode of "course cooperation+project cooperation +“IP co-creation", and 40 students will be recruited for computer, media art, news communication and other majors every academic year, and eight courses including Digital Man Making, Metauniverse and Media Philosophy, Introduction to Artificial Intelligence and Virtual IP Operation will be completed during the training period.

Laboratory scene

It is reported that half of the course content of the micro-major will be taught by the CDD team of Tencent Content Ecology Department in person, and students will have the opportunity to enter the actual combat of the enterprise in the name of project internship. In the teaching process, we also arranged visits and exchanges to broaden students’ horizons, hoping to transport innovative and creative talents that meet the needs of the intelligent audio-visual industry.

Wang Jianlei, the head of the micro-specialty of "virtual digital people" and an associate professor at the School of Communication, Shenzhen University, said that "virtual digital people" is a very cutting-edge professional direction, and Internet technology companies are groping for stones and finding ways to seize the new blue ocean of the market. In the education field, China Communication University, Nanjing University and South China University of Technology have all started to offer relevant virtual production courses consciously. The cooperation between Shenzhen University and Tencent Technology Co., Ltd. has inherent advantages. If we can take the lead in setting up a micro-specialty based on this, it will be the first virtual digital professional in China. With distinctive features and high focus, it will form a good professional recognition. The adoption of micro-professional training can deliver the shortage of talents to the market in the shortest time, which will certainly benefit all parties in Industry-University-Research.

IDC MarketScape： Baidu Security is the leader in NESaaS market.

2023年1月6日 | admin | dfgdfg

Recently, IDC, an authoritative international consulting organization, released the report IDC MarketScape: China Public Cloud Network Edge security-as-a-service (NESaaS) Market (hereinafter referred to as IDC Report). With its leading AI security technology capability, Baidu was in the leading position in the evaluation of China public cloud network edge security-as-a-service market in 2022.

IDC pointed out that based on the summarization and refinement of Baidu’s years of security practices, relying on artificial intelligence and big data, Baidu has terabyte (TB)-level cloud security protection capability under the all-round technical support and escort of Baidu’s security, and the risk of intercepting malicious web pages and user searches exceeds 100 billion times every year. At the same time, Baidu Intelligent Cloud Security is one of Baidu’s five security areas. Combining the technical advantages in AI field and the long-term service experience in cloud computing field, it has gradually established a complete and comprehensive multi-class security product matrix and security solutions to meet various business needs, forming a secure and reliable cloud base, effectively protecting the security of government and enterprise assets.

This is another recommendation after Baidu Security was selected as the recommended manufacturer of IDC’s "China Data Security Technology Development Roadmap, 2022" this year. It shows that Baidu Security has been at the leading level in the industry in AI security, data security, cloud security, business security and mobile security.

In recent years, with the increasing variety, quantity and importance of cloud assets, its allure to network attackers is also increasing. Based on the above network security requirements, IDC put forward the concept of Network Edge security-as-a-service (NESaaS) in 2022 in order to better guide enterprises to solve the network security construction in complex IT environment. IDC believes that NESaaS, as a stack of comprehensive network security capabilities, can naturally and efficiently integrate various security capabilities, thus fully embodying the concept of software-defined security access.

IDC’s report focuses on Baidu’s intelligent threat hunting platform, which is the product of Baidu’s integration of technologies such as deep learning and big data analysis. The platform has six advantages, such as powerful intelligent tracing, three-dimensional API protection, supporting 0day/Nday attack detection, automatic drainage honeypot, abundant attack and defense exercise toolbox, low false alarm and low false alarm, etc. It analyzes the network traffic of customers in the form of gateway, identifies the identity of attackers through powerful intelligent tracing mechanism, and assists customers in advanced security by combining real-time data and intelligent algorithm model.

Baidu’s intelligent threat hunting platform is based on big data intelligence and AI analysis model’s ability of accurate portrait and tracing of attackers, and realizes tracing analysis and blocking of users, equipment and network IP at three levels. It can identify the identity of potential attackers by association analysis combined with intelligence database, completely subverting the passive mode based on IP tracing in the past. Compared with the common honeynet tracing scheme in the industry, the tracing success rate is increased by 2~5 times, and the problem of low tracing success rate and easy circumvention in the industry is solved in a breakthrough.

At present, Baidu’s intelligent threat hunting platform has been widely used in energy, government affairs, transportation, Internet and other industries. With its outstanding AI security capability, it has made outstanding offensive and defensive performance and traceability performance in network offensive and defensive drills and red-blue confrontation actions for head customers such as government, enterprises and energy, and helped government and enterprise organizations win the first place in offensive and defensive drills.

In a project case of an energy company, by virtue of the product’s efficient perception and protection ability to unknown threats, it helps customers to deal with advanced threats such as 0day/Nday and APT attacks more efficiently, realizes rapid detection and response to security threats, and improves the efficiency of security operation and maintenance. It can not only build a complete defense-in-depth system, combine real-time data and artificial intelligence algorithm model, assist customers to make advanced security decisions, empower security operation and maintenance personnel, effectively deal with known and unknown threats, and fight against 0-Day and N-Day. It can identify the identity of attackers through a powerful intelligent traceability system, completely overturn the passive mode based on IP traceability in the past, and efficiently escort key infrastructure systems.

With Baidu’s leading AI security capability fully integrated into all modules of Baidu’s brain and covering all businesses in Baidu AI Cloud, under the unique advantage of "integration of cloud and intelligence", Baidu Security will continue to be innovative and pragmatic in the future, take on the strong demand for security from the market and enterprises, and devote itself to building an intelligent integrated security system based on AI and big data based on the rich technical precipitation of Baidu AI and big data, so as to accelerate the process of enterprise intelligent security integration.