AInnovation Ranks First in WIDER FACE, one of the World's Most Authoritative Face Detection Challenges

BEIJING, May 30, 2019 /PRNewswire/ -- Recently, WIDER FACE, one of the world's authoritative public evaluation dataset for face detection, released its latest rankings. "AInnoFace", the algorithm developed by AInnovation surpassed the performance of many well-known AI enterprises and university laboratories, including Baidu, Tencent, JD.com, DiDi, Carnegie Mellon, University of Chinese Academy of Sciences, among others. AInnoFace ranked first overall across all three subsets "Easy", "Medium" and "Hard" of the WIDER FACE Challenge under six evaluation criteria (ranking first in five criteria and second in only one).

Taking the "World's Largest Selfie" for example, the AInnoFace algorithm detected 918 faces with a high degree of detection accuracy. By comparison, Baidu's PyramidBox algorithm detected only 880 faces in the photo.

The AInnoFace algorithm of AInnovation is developed from the well-known RetinaNet single-stage detector, and adopts a IoU regression loss for Bounding-Box regression to improve location detection accuracy, implements a Selective Refinement Network for higher Recall Rates and lower False Positive Rates, utilizes an augmented strategy of rich data variety for more robust final detection model, leverages Max-out Label to make more accurate classification predictions, and uses an improved multi-scale detection strategy to detect better faces at different scales. Through the above mentioned improvements, the AInnoFace algorithm can maintain outstanding detection performance even at the extremes, if faces are small, dimmed or covered. Altogether, the AInnoFace algorithm is able to effectively improve the recall rate and accuracy of face detection in complex scenes, and addresses the technical problems of face detection in public scenarios.

About WIDER FACE

Face detection is an increasingly popular and challenging problem in the field of Computer Vision and one of the most important business scenarios for AI algorithms. In order to improve the performance of their AI algorithms and showcase their technical capabilities, many AI companies choose public datasets to validate their algorithms. Among the available face detection datasets, WIDER FACE is one of the largest and most challenging dataset. This dataset contains 32,203 images and 393,703 face tags, in which 40% are training images, 10% are validation images, and 50% are testing images. The data in each set can be classified as "Easy", "Medium" and "Hard" based on the difficulty of face detection.

Pictures in WIDER FACE are extremely difficult to recognize, and approach those in real scenarios. The dataset contains pictures with a variety of influencing factors, such as dramatic variation in face size, various photographic angles, changes in face pose, different degrees of occlusion and changes in facial expression, various types of light pollution, brightness considerations, and various makeup styles. Therefore, WIDER FACE is famous for its challenging datasets in the field, and attracts a large number of international technology giants, institutions and universities (including Carnegie Mellon and other established universities around the world, as well as famous AI companies such as Baidu, Tencent, JD.com, DiDi, IBM, and others).

About AInnovation and the Winner Team

Established in March 2018, AInnovation is an AI subsidiary of Sinovation Ventures. With the mission of "AI Empowering Business", it is committed to providing AI-related products and business solutions for enterprises using the most advanced AI technology. Hocking Xu, CEO of AInnovation, has more than 20 years' experience in sales, products, technology, services and management in the IT industry. He successively served as corporate executive to many of the world's top 500 technology giants including IBM, Microsoft and SAP, and has a deep understanding of the developments of various commercial fields in China. The AInnovation "two-wheel drive model" combines "technology products with industry scenarios", and gives rise to extremely fast commercialization.

There are many excellent R&D teams in AInnovation. The winning team of the WIDER FACE competition is composed of Faen Zhang, Xinyu Fan, Guo Ai, Jianfei Song, Yongqiang Qin and Jiahong Wu. The team leader Faen Zhang is currently the CTO of AInnovation, the Chief Architect of AI Engineering Institute of Sinovation Ventures, and Honorary Professor of The University of Nottingham Ningbo China. He had worked at Microsoft, Google and Baidu, and made great achievements in the field of AI industry and academia. He holds many international AI patents and has published several Top-level Conference Papers on AI. Other members of the team also have deep education backgrounds and practical experience with AI.

Since its founding, AInnovation has strived to develop commercial products using high-quality AI algorithms for the retail, manufacturing, finance and other industries. It focuses on the advancement and maturity of AI algorithms, and provides various commercial AI products and solutions. As a result, the company has developed at a rapid and steady pace. At present, AInnovation has established an advanced AI data platform and self-developed deep learning support platforms. Furthermore, it has produced several Top-level Conference Papers which improve the accuracy of the algorithms in the business scenarios improve the speed of model training, and integrate computing resources very efficiently.

SOURCE AInnovation