YAZAKI Image Annotation Service

Service Customer case study

For over 80 years since our founding, YAZAKI has engaged in "monozukuri" (Japanese-style manufacturing), becoming a company with a top-class share in the wire harness (automotive parts) industry through the teamwork of our employees.By combining the insights of people who have driven and supported technology with AI development, we offer AI development support and high-quality, unparalleled annotation services.Our annotation service supports economies of AI engineer resources through high quality and speedy delivery.

Overview

YAZAKI Image Annotation Service realizes high quality annotation in the three processes of "specification review," "annotation work," and "data delivery," which determines the quality.

Specification review: A project manager with AI knowledge and experience proposes the definition of requirements and prepares specifications to meet the client's AI development requirements.
Annotation work: The actual work is performed by a site director with operational know-how honed in the automotive industry and dedicated, qualified workers who create accurate data.
Data delivery: Triple checks are performed to prevent human error and ensure the delivery of high-quality annotation data.

We are choosed by following customers

Case1

Engineer resources related issue
- With limited engineering resources, time is taken up by annotation work, making it difficult to concentrate on development work.
- We outsourced a complex annotation project, but the requirements were not communicated correctly, resulting in confirmation and revision reviews.
Case2

Quality related issue
- The quality of the annotation data was poor, which adversely affected the accuracy of the AI model.
- Many revisions of annotations were exchanged, resulting in delays in the schedule.

Features of YAZAKI's image annotation service

01.

Delivery of high-quality data

By employing professional annotators and performing triple checks, we achieve high-quality results.

02.

Flexible handling of highly difficult projects that other companies cannot handle

We can handle even unclear images that have a large number of target objects, such as satellite images in a flexible and careful manner. Even with ambiguous requirement definitions, workers are able to create data of high quality with little variations through face-to-face interaction.

03.

Proposals based on practical AI development expertise

We are able to make practical and unique proposals based on the extensive know-how we have gained by being in charge of AI development for the YAZAKI Group and leading the development of products using AI and the introduction of AI.

Examples of supported industries

Satellite & drone images

Map creation: We support satellite images, which are said to be extremely difficult because of the large number of objects that appear in one image.

Factories & robots

Picking robots: Using our expertise cultivated as a manufacturer, we can handle detailed rule settings.

Automotive & smart cities

Self-driving: We can handle complex images, such as rainy or nighttime images and intersections with numerous target objects.

Agriculture, forestry & fisheries

Agricultural robots: YAZAKI can handle all processes, starting from rule setting, even for natural objects with individual differences such as produce, for which it is difficult to set up annotation rules.

Customer case study

SkymatiX, Inc.

The challenge of handling large volumes of complex annotation data in-house was solved by outsourcing to our service in a way that maintained quality and speed.
The data we deliver is used for AI development by Skymatics Inc.
Here are some examples of how we have been used.

Use cases

Service use procedure

We will provide an estimate within two business days at the shortest after the requirements interview.

STEP1　Service Introduction

After you submit an inquiry through the form, we will have a meeting to introduce service details and examples.
STEP2　Requirements interview

We will also ask about your company's AI development overview and annotation requirements.
STEP3　Estimate

We will present an estimate after discussing requirements, delivery formats, and deadlines.
STEP4　Execution of contract/finalization of requirements

We will then sign the contract and NDA.
STEP5　Annotation implementation

You will send us the necessary data. We will conduct QA as appropriate to ensure annotation based on a correct mutual understanding.
STEP6　Delivery and acceptance inspection

We will deliver images that have been quality checked by us. Following inspection by your company, the project will be completed.