Visual Edge-AI for Division
4DIViSiON will always strive to impress customers.
Our Product
home > product
TesseraVue (2ch)
It is a 3D sensor that can observe various surfaces and measure them as standard. The image taken by the camera and the start is analyzed with a self-developed 3D algorithm to generate point cloud data. In 2CH, one more camera was added (total of two) to increase the recency of 3D data. By providing 2D images and 3D data at the same time, overpainting, blank picking, and coloring can be realized.
main features
- Scan various materials
- Fast scan time
- High accuracy
- Provides two versions of NIR band and visible light band
TesseraVue (NIR)
It is a 3D sensor that can measure the surface of various materials effectively and in real time. Point cloud data is created by analyzing images taken using a single camera and projector with a self-developed 3D algorithm. Since 2D image and 3D data are provided at the same time, the best bin-picking solution can be realized.
main features
- Scan various materials
- Fast scan time
- High accuracy
- Provides two versions of NIR band and visible light band
WrinkleVue
WrinkleVue scans and analyzes your face in real time. It generates 3D data of the face and analyzes the skin with an advanced algorithm developed in-house using three cameras and a projector. Outputs quantified data on the pigment, wrinkles, and pores of the user's skin. This data can help tailor solutions for your skin care.
main features
- Generate 3D data of face
- Convenient scanning with high-speed shooting
- Quantified data for skin care
RoboVue (2D Sensor)
A device that quickly and accurately inspects the surface of railway rails.
Analyze 2D images taken with line scan to find corrugated wear.
It analyzes even the parts that people might miss in detail and accurately.
main features
- Automated analysis of railway surface
- High accuracy of 96%
NVIDIA Jetson Carrier Board
The jetson tx2 carrier board is designed to accommodate NVIDIA's Jetson tx2 series (TX2, TX2 4GB). Modules with ARM CPU and NVIDIA GPU provide high specifications for AI and deep learning even in embedded systems. The design of the board includes USB3.0 x2, micro USB2.0 x1, HDMI x1, GbE x1, UART x1, and GPIO 20 pins. It comes with Ubuntu OS for users to get started easily.
main features
- High performance AI and deep learning
- Equipped with ARM CPU and NVIDIA GPU
AI Face Recognition Solution
The AI face recognition solution uses cameras and various imaging devices to recognize human characteristics. With real-time monitoring and accurate face recognition, it can be used not only for services such as access control, time and attendance management, detection of specific people, and credit authentication, but also in the retail field as demand for unmanned stores increases. 4DiViSiON is an AI deep learning-based face detection and 1:1 / 1:N face recognition algorithm that enables high-speed / high-precision detection despite various constraints in reality such as face angle, glasses, and makeup. Using a near-infrared image based fake face attack detection algorithm, it blocks attempts to recognize non-real people, such as high-definition photos, 3D models, and human videos. Along with the face recognition solution, a non-contact fever check solution is provided at the same time according to the customer's situation, and 4DiViSiON's unique technology has improved the accuracy of remote body temperature recognition and body temperature measurement.
AI machine vision solution
AI machine vision solutions are the best deep learning-based vision solutions designed for factory automation.
It is designed based on field-tested, optimized and proven technology and machine learning algorithms.
You can identify product defect information from image information and classify it by type for efficient defect management.
4DiViSiON's 2D machine vision solution acquires images in various resolutions to analyze and detect product defects. It can detect and classify defects such as PCB boards, packaging paper surfaces, and machined products with specific textures.
4DiViSiON's 3D machine vision solution analyzes images taken using cameras and projectors with self-developed 3D algorithms to generate Point Cloud Data to analyze depth and direction information that are difficult to analyze. Real-time 3D shape inspection is possible by effectively measuring the surface of various materials.
Steel Defect Detection Image Segmentation using U-net
Segmentation and detection of defective parts using deep learning in metal surface images
Different type of metal defects
AI Linescan Defect Analysis Solution
By taking a picture of the rail with a linescan, it is possible to analyze the values of the wave length and the start and end points of the wear, giving an idea of the size of the wave wear. The AI line scan defect analysis solution can be applied not only to rails, but also to roads, bridges, and everywhere where corrugated wear can occur.
Rail corrugation automatic recognition
네이버 아이디로 로그인 
카카오 아이디로 로그인 
구글 아이디로 로그인 