How do companies that are informational in the park have the ability to “make blood” while precipitating data?
In the park’s informational track, everyone’s cut-in path is mainly based on the S2B2C model, which focuses on B-side information services, such as the previously reported Pineapple School Bus, Morning Bud Education.
The recently contacted “Red Orange Cloud” focuses on face recognition technology and provides kindergarten teachers with services for photo uploading, identification, and distribution to parents. The specific process is for the teacher to take photos of the child in the teaching process through the Red Orange Cloud APP. The back end is automatically uploaded, and the classification (including characters, behaviors, and works) is distributed to the parents through face recognition technology. Parents can receive it in real time on the small program side.
In the early stage, Red Orange Cloud was mainly profitable through the production of “Growth Album”. The exquisite version of the growth album is currently a fee-based product in some parks. The unit price of single-item customers is around RMB 200, and the third-party and kindergarten adopt a cooperative sharing model. The late Red Orange Cloud plans to cut into the core teaching services, based on mature face recognition technology + child growth ability model + large amount of data, to personalize the child’s growth and guide the kindergarten teacher’s teaching process.
The profitability of the growth album is only the goal of the early stage of the Red Orange Cloud. In the future, I hope to find a way to obtain sustainable and effective child growth data through this portal, and to personalize the child based on the large amount of data.
The business logic of Red Orange Cloud is:
1)Growing photo albums have become the basic service of kindergarten, but need to be improved
In the past, every time a kindergarten graduated from a large class, the kindergarten will work with some studios to take photos of the graduation. With the emergence of home-grown products such as Jiatong Home, Smart Tree, etc., the teacher can record the photos of the children in the kindergarten to the parents at any time, so the child’s “growth album” in the kindergarten has become the basic service provided to the parents.
He Zhiyong, founder of Red Orange Cloud, told In the past, there were 12,000 photo studios serving kindergartens each year. The market volume of the graduation book is 3 to 5 billion. If it is dismantled to semester, it will reach about 20 billion yuan. The size of the market around.
The existing growth photo album on the market is still in a rather cumbersome state in the process: teacher shooting – teacher uploading public community / private sending to parents – parents looking for photos of children in the community – self-storing – one-click generation of a certain capacity Digital Growth Album – Print as intended, third party is responsible for sending.
Partial recognition technology part analysis process
The process of combining face recognition technology is: teacher shooting – parents receive photos classified by label – platform automatically selects pictures related to children’s teaching to generate albums – print as intended, third party is responsible for sending.
Integrated, it is to reduce the workload of teachers and improve the quality and efficiency of photo collection. At present, Red Orange Cloud has cooperated with more than 200 kindergartens, covering more than 40,000 children, and accumulated more than 2.4 million photos. It is planned to expand to 1,000 kindergartens in 19 years.
2) The current data for children is missing, and the longer-term goal is to build a children’s learning and evaluation system with the advantages of “data and technology”
Compared with the K12 field, there is no feasible commercial realization mode in the teaching scene during the age of 3-6, and the recognition threshold of the child is much higher than that of the adult. The data of the child is still in a missing state.
But in the long run, He Zhiyong said that the predicament of the kindergarten side has always existed: the quality of the teacher needs to be improved, and the awareness of self-growth is not high enough, and the supply of high-quality teachers is lacking. Effective data accumulation and child-based ability algorithm models, as well as mature technical support, can reverse the teacher’s teaching and lower the teacher’s teaching threshold.
The current database and technology level still cannot be realized. He Zhiyong believes that it will take at least 9 years to realize this scenario, but once realized, its value will be huge. But from the current point of view, Red Orange Cloud needs to have the ability to “hematopoietic” by growing a photo album, or other monetization modes of the platform, while continuing to deepen the technology and “nurturing” data. How to weigh the data and realize the value is a big test for the red orange cloud.
He Zhiyong said that the core strength of Red Orange Cloud lies in children’s face recognition technology:
They have done a comparison, using the industry’s largest and widely used Microsoft Million Face Dataset + InsightFace Model Framework as the standard model, with 99.99% correctly identified, Red Orange Cloud’s child face recognition model There is a 10% improvement in accuracy compared to the standard model.
With algorithms and data advantages, Red Orange Cloud also has edge computing capabilities that increase transportation efficiency (20-50ms of computing time) and protect data privacy.
Team, R&D and algorithm personnel have experienced in the field of face recognition. The expert team was the head of Tencent’s deep learning platform/wide point. The person in charge of teaching and research has 12 years of curriculum development experience, which is an international child development and Specialist for the Education Society.
Red Orange Cloud has launched the Pre-A round of financing.