Next-generation smart home AI

Next-generation smart home AI

Technology News |
By Jean-Pierre Joosting

IMOU, a leading smart IoT provider, has launched IMOU SENSE, its self-developed AI algorithm matrix, with aims to create a more convenient and efficient experience for smart product users. The IMOU SENSE algorithm matrix consists of intelligent, navigation and perception algorithms.

The intelligent algorithms include video AI technologies, involving attribute recognition (family member attribute recognition, gesture recognition, amongst others), target detection (human figure, vehicle, package, face, pet detection, and so on), target tracking, behavior analysis (fall detection, perimeter intrusion), and audio AI technology involving 3A algorithm, sound source localization, abnormal sound detection, amongst others.

The navigation algorithms, mainly used for home service and cleaning robots, consist of the SLAM algorithm, the path planning algorithm, and the decision-making control algorithm. The perception algorithms are mainly responsible for calibrating multi-source sensors and multi-source sensor information processing and fusion. The sensors involved include binocular vision, structured light camera, millimeter wave radar, line laser, TOF camera, etc. Typical applications include scene depth estimation, liveness detection, robot obstacle avoidance, etc.

IMOU algorithms have made breakthroughs in many fields, such as object detection, object tracking, change detection, and cutting-edge technical pre-research. These achievements include VOT-ST2020, winner of an international competition on visual object tracking, Best Session Paper of the 2nd Asian Conference on Artificial Intelligence Technology, first place in the comprehensive accuracy ranking of the generic object tracking benchmark (GOT-10k), and first place in the large-scale dataset for object detection in aerial images (DOTA). 

Depending on the cloud platform, IMOU SENSE obtains user data, improves the quality of training data, and realizes a data closed loop, which, on the one hand, can continuously improve the accuracy of the algorithms, and on the other hand, realizes the generalization of the algorithms to various home scenarios. IMOU SENSE is customized and optimized at the chip level so that the AI chip can be used in the best way possible.

In actual smart home scenarios, IMOU SENSE AI supports accurate detection of multiple target categories, such as human, vehicle, pet, face, and package, and supports any arrangement and combination of the above target categories. In addition, with the help of the target tracking algorithm, the camera equipped with IMOU SENSE can continuously follow the target of interest, record the trajectory, and report the information in time when abnormal behavior is detected.  IMOU SENSE enables real-time detection and feedback, with an accurate detection result that is 20-to 50-percent higher in accuracy rate than the algorithm that accompanies a chip. 

IMOU SENSE has a deep understanding of scenarios, which makes it really smart. For cleaning robots equipped with IMOU SENSE, the robot will not only know where it is, but also know what is around itself, and how it can work better. At the same time, IMOU SENSE can adapt to unknown and unstructured scenarios. When the robot is placed in an unfamiliar environment, it has the ability to independently explore and learn the surrounding environment.

Using a cleaning robot equipped with IMOU SENSE for example, simply makes the cleaning process smarter, with IMOU SENSE guaranteeing the detection of key frame dynamic updates, allowing the robot to better understand the scenario and deal with unknown situations. It can also complete map building and route planning quicker and more accurately when dealing with unfamiliar environments and obstacle avoidance, to make cleaning worry-free.

The cleaning plan is also smarter. The high-precision positioning and fusion mapping capacities of IMOU SENSE AI support more customized cleaning plans to adapt to various situations in the user’s home, making the cleaning more efficient.

IMOU SENSE supports the detection of more than 30 object categories of home scenarios, and designs customized obstacle avoidance strategies for each category to ensure the safe operation of cleaning robots. In addition, relying on the scene recognition algorithm, the IMOU robot can autonomously identify the scenario it is in, and then provide customized and personalized cleaning solutions. For example, when it is detected that the robot is in the kitchen or bathroom environment, it will automatically increase the mopping strength; when it is detected that the robot is in the bedroom environment, it will dynamically adjust the suction to avoid introducing excessive noise.

The newly launched security cameras Cruiser 2 and Rex 3D are the first IMOU products to be equipped with IMOU SENSE. More products are to come with IMOU SENSE, providing diversified AI solutions for different usage scenarios.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles