Ukrainian hardware giant Ajax Systems is training its own neural network for its IndoorCam AI cameras. Ajax Systems introduced the IndoorCam AI camera at its Special Event product presentation last November.
Ukrainian hardware giant Ajax Systems is training its own neural network for its IndoorCam AI cameras. Ajax Systems introduced the IndoorCam AI camera at its Special Event product presentation last November.
As the company’s CMO Valentyn Hrytsenko told Forbes, the development of a neural network for IndoorCam AI cameras began in 2022. The company, according to him, is still catching up with the market. «AI in cameras and doorbells is the standard on the market,» says Hrytsenko. «A person cannot simultaneously and qualitatively view a video stream from a location where, for example, 150 cameras are installed.»
Hrytsenko does not disclose investments in development.
Artificial intelligence identifies objects in the video stream that it considers dangerous and gives a signal to a person. Then it reviews a specific part of the video and checks whether it was a false alarm or not, says CMO Ajax Systems. «On average, up to 20 cameras are ordered per object. This is the home segment and SMEs,» says Hrytsenko.
The company trains the neuron on data sets it receives from users. «We have a beta testing program, where participants independently give us access to individual videos,» says Hrytsenko.
These videos are then analyzed by the company’s specialists. «If the AI detects an animal in the video, but when viewed, it turns out that it is a branch, we additionally 'feed' this clip to the neural network for further training,» says Hrytsenko.
Currently, says Gritsenko, the demand for such cameras is obvious. In four weeks, pre-orders in Australia for the IndoorCam AI camera have offset the costs of the global Ajax Special Event product presentation. Australia is currently, he says, the largest market in terms of pre-orders for this product, followed by the Benelux.
Currently, the company’s largest share is generated by intrusion protection products, Hrytsenko admits. «Video products will be able to beat this trend in three to four years, but we need to expand the line,» he says.
Among other plans for the Ajax Systems video surveillance division is to add color images to cameras at night. The quality of the image depends on the resolution of the matrix in the camera, he adds. «The better it is, the more expensive the camera itself will be,» says Denisenko.
Ajax Systems does not disclose how much this feature might cost. The plans for this year are to increase revenue from video surveillance to 20-25% of the total. «If it works out, it will be a good story,» says the CEO of Ajax Systems.
