How Data Annotation Services Fuel Self-Driving Industry 2021-Part5
“In the first half, whatever is good will be installed. In the second half, we have to consider the supply chain, cost, durability, adaptability, and other issues.” Peng Jun, co-founder, and CEO of Pony. AI
As the second half of the competition is from 1 to N, there are many uncertainties. And the focus of technology companies and car companies is not the same.
2021 is known as the first year of intelligent vehicles, and both automatic driving and AI- assistant driving are given attention. It is generally believed in the industry that the first half of automatic driving is to prove the feasibility of the technology. In the second half, the competition of cost control, scale, and operation ability are more important.
With the expansion of the application of Robotaxi (driverless taxi), all technology companies are moving forward to reduce costs, conquer mass production and expand the scope of operation; With the deployment of laser radar and other sensors, car companies are committed to promoting more multi-driving function of the car.
Technology companies value technology, while auto companies value profitability. It is said that Chinese vehicle manufacturing is in the lowest value part, while automatic driving software, related supporting communication operation services, high-precision map services, and travel services are in the higher value part. Traditional car companies, new carmakers, and technology companies will not miss this blue ocean.
However, regarding the landing of high-level automatic driving, car companies, and technology companies should not be too optimistic. 2021 is also known as the first year of mass production of high-level automatic driving. Mr.XU, executive vice president of Bosch China said, “there are many demos for L4, but there is still a long way to go before achieving mass production due to the limitation of technology and regulations.”
The World’s First Driverless Law to Allow L4 Class Automatic Driving on the Road in Germany
“In terms of the current landing situation, it is mainly on the L4 experience and L2 responsibility. The driver should still be ready to take over the vehicle at all times.” Mr. Zhang, vice president of Horizon.AI and general manager of intelligent driving product line, said “only with the improvement of chip computing power, technology companies can cover more and more scenarios, including an unprotected left turn in complex urban scenes.”
The goal of the second half has been settled, that is, to improve the capacity of large-scale mass production and commercial profitability, and to “feed” technology with massive data. For technology companies, from 0 to 1 has been completed, but they still need to reduce costs and increase efficiency and move towards the larger-scale operation. For car companies, how to break the L3 paradox and improve the customer’s driving experience are the main focus.
Many autopilot technology companies have a consensus: autopilot has completed the verification from 0 to 1 in the first half match, that is, the demo in the laboratory has been implemented on a small scale, and the technology is feasible.
Waymo, a driverless company hatched by Google, and Pony.AI, a domestic driverless company, have completed the operation from laboratory to small-scale landing operation, while another company Autox is committed to removing safety officers. In the beginning, the car companies which put the cost and commercial implementation in the first place have also provided solutions for advanced assisted driving in limited scenarios through self -developed team or cooperation.
Due to different strategic routes, although they are all in the transition stage, there are still differences in the dividing line between the first half and the second half among many technology companies.
In the view of the global general manager of Momenta.AI, an automatic driving technology company, “the first half and the second half” is a virtual reference, not a “middle point” for technology or commercial application.
Compared with other players, beginning from L4 technology, Momenta adapts the mass production data-driven “flywheel mode” at the beginning of its establishment, which leads to the “two directions” product strategy — mass production automatic driving and completely unmanned driving.
It’s very difficult to push the flywheel at the beginning, and it takes a lot of effort to build a data-driven closed loop. However, if we continue to keep using data to push the flywheel, it will turn faster and faster, and the growth curve is not linear, but explosive.
In conclusion, the player who can first achieve from 1 to N, with the large-scale ability and large-scale commercial demonstration operation, can stand out in the second half of the competition.
“The speed of a company’s progress depends on the distance between products and commercial applications, which is the only measure of autonomous driving in the second half.”
Last year, at the CVPR 2019, Andrej Karpathy, the Senior Director of AI at Tesla responded to the question below:
how to estimate the volume of labeled data required to train and validate the self-driving cars for a particular scenario?”
378 hours of data.
The more accurate annotation is, the better algorithm performance will be.
Any tiny error during a driving experience may lead to dreadful results. Nowadays, people are more and more concerned about the driving safety issue as several self-driving automobile accidents happened.
For technology-type car companies, from 0 to 1 has been completed, but they still need to reduce costs and increase efficiency and move towards larger-scale production.
With the tremendous amount of training data and the high accuracy requirement, a high-quality data annotation service is crucial to guarantee autonomous vehicles are safe for the public.
Common Data Labeling Types Include:
It’s challenging for self-driving manufacturers to internally meet the burgeoning demand for high-quality data annotation.
The labeler used to annotate point by point, which cost lots of time.
3D annotation and video annotation are considered the toughest services in data labeling. At present, object tracking algorithms based on machine learning have already assisted video annotation. The annotator annotates the objects on the first frame, and then the algorithm tracks the ones in the subsequent frames. The annotator only needs to adjust the annotation when the algorithm doesn’t function well. It is 100 times faster than before.
Thanks to an AI-assisted system, the corresponding parts can be automatic can be transcribed and the human only needs to check and modify the wrong parts.
Nowadays, some AI-assisted tools come to practice, standing out in 2 factors.
- Cost reducing: With the help of AI-assisted capabilities, clients can save more money as the labor cost goes down.
- Time reducing: Make the large-scale requirement of training data done in a short time. Using AI-assisted tool can improve efficiency multiple times
Can we get rid of the human workforce?
The answer is no.
In fact, manually labeled data is less prone to errors regarding quality assurance and data exceptions.
The human workforce cannot be totally replaced by some tools leading with an AI-based automation feature, especially dealing with exception, edge cases, complex data labeling scenarios, etc.
ByteBridge is a human-powered and ML-powered data labeling tooling platform. We provide scalable, high-quality training data for ML/AI industry with flexible workflow.
Quality and Accuracy
- ML-assisted capacity can help reduce human errors by automatically pre-labeling
- The real-time QA and QC are integrated into the labeling workflow as the consensus mechanism is introduced to ensure accuracy.
- Consensus — Assign the same task to several workers, and the correct answer is the one that comes back from the majority output.
- All work results are completely screened and inspected by the machine and human workforce.
In this way, ByteBridge can affirm our data acceptance and accuracy rate is over 98%.
- Progress preview: client can monitor the labeling progress in real-time on the dashboard
- Result preview: client can get the results in real-time on the dashboard
- Real-time Outputs: client can get real-time output results through API. We support JSON, XML, CSV, etc., and we can provide customizable datatype to meet your needs.
3D Point Cloud Annotation Service
ByteBridge self-developed 3D Point Cloud labeling, quality inspection tool, and pre-labeling functions can complete high-quality and high-precision 3D point cloud annotation for 2D-3D fusion or 3D images provided by different manufacturers and equipment, and provide one-station management service of labeling, QA, and QC.
More info: ByteBridge Launches World’s First Mobile 3D Point Cloud Data Labeling Service
3D Point Cloud Annotation Types
- Sensor Fusion Cuboids: 12 categories include car, truck, heavy vehicle, two-wheeled vehicle, pedestrian, etc.
- Sensor Fusion Segmentation: obstacles classification, different types of lanes differentiation
- Sensor Fusion Cuboids Tracking
① Tracking the same object with the same ID, labeling the leaving state;
② Point clouds or time-aligned images could be provided, point clouds outputs only.
A collaboration of the human-work force and AI algorithms ensure a 50% lower price compared to the conventional market.
As the quality of the labeled dataset determines the success of the self-driving industry, cooperation with a reliable partner can help developers to overcome the data labeling challenges.
We can provide personalized annotation tools and services according to customer requirements.
If you need data labeling and collection services, please have a look at bytebridge.io, the clear pricing is available.
Please feel free to contact us: firstname.lastname@example.org
1 Didi Autopilot Has Completed a New Round of Financing of over US $300 million
2 Which One is More Important to Automatic Driving, Algorithm, or Computing Power?
3 Is Lidar the Future of the Self-Driving Industry2021?