ADAS Traffic Labeling

Automated driving systems use vision, radar, ultrasound, and combinations of sensor technologies to automate dynamic driving tasks. These tasks include steering, braking, and acceleration. Automated driving spans a wide range of automation levels - from advanced driver assistance systems (ADAS) to fully autonomous driving.
- Experience in the tens of thousands of hours of work on annotation frames
- Accurate labeling by frames

Scene Labeling

Scene Labeling and Re-Simulation Automated driving record large amounts of sensor data from test vehicles including video, radar returns, and other information on the state of the vehicle. We use a process called ground-truth labeling to annotate recorded sensor data with the expected state of the automated driving system. They can be performed to open-loop testing of the automated driving features by processing recorded sensor data and comparing the system output to the human-verified ground truth.


The functionality of the object recognition application can be verified with applicable software tools. Unfortunately, the automatic verification malfunctions in hazardous weather conditions and with poorly recognizable objects. These specific objects must be evaluated manually. In these specific cases it is necessary for a person decide which objects are relevant or non-relevant for the ADAS.


Autonomous driving is a key factor for future mobility. Properly perceiving the environment of the vehicles is essential for a safe driving, which requires computing accurate geometric and semantic information in real-time. In this paper, we challenge state-of-the-art computer vision algorithms for building a perception system for autonomous driving. An inherent drawback in the computation of visual semantics is the trade-off between accuracy and computational cost. We propose to circumvent this problem by following an offline-online strategy. During the offline stage dense 3D semantic maps are created. In the online stage the current driving area is recognized in the maps via a re-localization process, which allows to retrieve the pre-computed accurate semantics and 3D geometry in real-time. Then, detecting the dynamic obstacles we obtain a rich understanding of the current scene. We evaluate quantitatively our proposal in the dataset and discuss the related open challenges for the computer vision community.

Labeling process

In this context, our Company is focused on combining computer vision techniques as pattern recognition, feature extraction, learning, tracking, 3D vision, etc. for developing in real-time algorithms that are able to assist the driving activity in future ADAS technology.


Service solutions

Our solutions are focused on high-quality data, automation of labeling process, optimization of costs and time expenditure as well as on valuable technical consultancy.

Precise Reporting of:
- Project progress
- Labeling accuracy
- Team statistics
- Customer specific KPIs
- Data Labeling
- Data management
- Labeling
- Quality control