About Us
Integrate Machine Learning for Organizational Needs
Neirolis is an International group founded by a team of deep machine learning scientists, machine vision engineers and excellent software developers.
We have a diversified background of artificial intelligence (AI) system development, Industrial 4.0 solutions, biometric identification and media analysis solutions for public security applications, media agencies, retail and marketing.
We like to invent algorithms, which work in unconstrained real-life scenarios.
Detection. Recognition. Analysis.
Data value with video analytics
Products
We are focused on empowering both people and organizations, by solving the most pressing challenges. We infuse video analytics into everything we deliver across our computing platforms and experiences.
Face Recognition
Face recognition video analytic transforms what is possible with a video surveillance system to identify people in photos, video, or in real-time. The systems use computer algorithms to accurately pick out specific, distinctive details about a person's face.
Emotion Analysis
Fast and accurate recognizing 6 emotions detection: Happiness, Sadness, Surprise, Anger, Fear, and Disgust. It can detect neutral face, Opened and closed eyes. These emotions are understood universally with particular facial expressions.
Object Detection
Object detection algorithms typically use extracted features and learning algorithms to recognize instances of an object category. More than 20000 object classes recognition including: people, vehicles, guns, helmets, attire, furniture, appliance, electronics and so on.
Object Tracking
Taking an initial set of object detections and create unique ID for each initial detection then tracking each of the objects as they move around frames in a video while maintaining the assignment of unique IDs. Detecting and recording tracks of vehicles, people or objects through the multiple cameras or video streams.
Scenes & Activities Analysis
The main goal of video analytics is scene analytics qualifies the motion as an object, understands the context around the object, and is able to track the object through the scene. Detection of people crowding, flying drones, fire or smoke, crossing lines, abandon bags and many other throughout the day and nighttime.
Text Recognition
Text recognition led with accuracies while still being computationally efficient. We train it by cast it as a sequence prediction problem, where the input is the image containing the text to be recognized and the output is the sequence of characters in the word image. Extraction and recognition of the text from car plates, signs, badges, video streams in more than 200 fonts and 164 languages.
Use Cases
FEATURES
-
State-of-the-Art Algorithms
RTMIP uses the bunch of neural networks, trained on more than 10 billions of unique images. Its high accuracy is confirmed by independent tests and contests. Moreover, we pay great attention to the performance of our algorithms without compromising the accuracy, sharpening the neural network for a specific task, which allows us to be ahead of the competition. We are always on the bleeding edge of machine learning trends, constantly updating our neural networks and making them faster, more accurate and relevant.
-
Highest Accuracy
Our unique face recognition neural network operates with an error rate as low as 1 in 1,000,000. With a test dataset of 10.000 unique people, its identification accuracy rate is up to 99.7% and its biometrics verification accuracy rate is up to 99.9%, which means the system can be used for every face recognition task.
-
Highest Speed
Our networks runs up to 100 times faster than competitive neural networks without losing accuracy. We pay attention to constantly upgrading our networks to run even faster.
-
Fastest Large Scale Search
Our face recognition algorithm handles millions of different faces. The RTMIP unique proprietary index engine is able search through billions of photos in less than a 0.1 sec. Face detection takes from 200 ns. Extraction of biometric personal characteristics takes about 500 nanos.
-
On Your Hardware
Our software runs with all data stored on your own servers to give your business absolute peace of mind. With cross-platform REST API on board and stack of native APIs, the RTMIP server can be easily integrated in any web, standalone mobile, or desktop application in a prompt and unified manner.