Intelligent Pet Monitoring Revolution: Lingyu Technology Aims to Be the AI Nanny for Fur Babies | Early Project

Author | Ye Danxuan

Editor | Yuan Silai

With AI being increasingly widely applied in human daily life, many traditional hardware industries are actively embracing the transformation of algorithms, and pet monitoring is also among them.

For contemporary people, pets have long gone beyond the category of domesticated animals and have become "furry children" and family members who accompany each other. Pet owners are also more willing to invest energy and money in the care and daily monitoring of pets: It is becoming a common phenomenon for office workers to install monitors at home and check the condition of their pets through the camera and have a remote conversation with them during the time they are away from home.

Lingyu Technology, which Hardcore recently came into contact with, has launched a pet camera - SiiPet - that can independently record the wonderful moments of pets and analyze their abnormal behaviors by integrating its self-developed AI visual algorithm system.

The founders of Lingyu Technology, Fu Xingyu and Wan Weitao, have rich experience in intelligent hardware and algorithm development. Based on the observation of pet families and the market, they found that the traditional pet monitors on the market currently have unsatisfactory effects in very important characteristics in vertical scenarios such as high-speed tracking and behavior understanding. There are still relatively significant technical pain points that require more in-depth AI algorithms and data accumulation to break through.

A common annoyance for pet owners who are office workers is that the time they really spend with their pets is only a few hours after work, and the time for pet owners who often work overtime or travel on business is even less. During the rest of the time, "furry children" can only live alone at home, and a large amount of pet life content cannot be discovered and recorded. More importantly, even if pets show abnormal behaviors related to health, it is difficult for pet owners to detect them through traditional monitoring and short-term interaction.

Based on the thinking about this problem, Lingyu Technology embeds a series of self-developed AI visual algorithms into the SiiPet pet monitoring camera. Through the core edge-side algorithm and cloud big model tools, combined with the underlying big data sets such as pet real behavior data, clinical data, and aesthetic evaluation data, it realizes the automatic capture of pet highlight pictures, abnormal behavior alerts, and multi-pet recognition.

In the specific use process, SiiPet, through the 4k camera equipped with an AI vision system, can accurately identify and ultra-high-speed track pets that are moving at high speed. At the same time, the camera will automatically capture and compose the highlight pictures to ensure that the pictures are as consistent as possible with human photography aesthetics, which is different from the traditional monitoring pictures.

Not only that, SiiPet can currently accurately perform multi-pet recognition, and archive and push the highlight clips of different pets in the background. Pet owners can switch to view the status of different pets in real-time and share the daily highlight pictures and short videos on their favorite social media with one click.

It is worth noting that the unique abnormal behavior capture function of SiiPet fills the gap in the pet behavior and health monitoring in the current pet camera field.

Behaviors such as vomiting, rubbing the buttocks, scratching, shaking the head, limping, or convulsing and their frequency of occurrence are important indicators for judging the abnormal health status of pets. However, even if a large number of pet owners install the existing pet monitors, they cannot accurately monitor the above behaviors and the corresponding frequency of occurrence. As a result, a large number of early health problems are ignored, and pets are also likely to miss the best treatment window.

Currently, the abnormal behavior capture function of SiiPet can be automatically monitored and captured through its intelligent system, continuously learned and accurately determined based on the built-in veterinary ethology knowledge and background user data, and pushed to pet owners.

Fu Xingyu told Hardcore that since the hardware supply chain in the current pet supplies market is already very mature, the industry is entering a stage of "involution" in terms of hardware indicators and costs, and it urgently needs to overcome and solve the existing problems through technological innovation.

"The tracking, recognition, and analysis capabilities of the camera are essentially the intelligent differences brought by the algorithm. I firmly believe that the development of pet supplies from networking and automation to true intelligence is a clear trend," Fu Xingyu said. He noticed that some pet companies are recently seeking algorithm support and introducing more AI technologies, and Lingyu Technology is establishing cooperation with them.

In terms of the team, the core technical team of Lingyu Technology comes from major intelligent hardware manufacturers. The CEO, Fu Xingyu, is a master's degree from the Department of Electronic Engineering at Tsinghua University. He once served as a senior product manager at tp-link and Xiaomi, responsible for multiple smart home products with an income of over 100 million. The CTO, Wan Weitao, is a doctor from the Department of Electronic Engineering at Tsinghua University. He once served as the person in charge of WeChat's general object detection and the AI visual capabilities of mini-programs, and has published multiple top AI conference papers.

Provided by 36Kr

Provided by 36Kr