Skip to content
联系我们

DaoAI World

Adapting to Diverse Industrial Automation Applications

2018-1.001

Leveraging the latest AI technology with 9 models. DaoAI World delivers powerful AI-driven machine vision tools directly to your fingertips. It achieves an "all-in-one learning" approach, minimizing human efforts in maintain system learning.

Screenshot 2025-02-12 at 5.19.19 PM-2
ALL IN ONE PLATFORM

Enable inspections from basic to advanced

Screenshot 2025-02-13 at 3.47.45 PM
Leveraging 9

Well-Tuned AI Models

DaoAI World provides 9 fine-tuned AI models that have been verified across tons of real-world in-field applications, offering numerous advantages, including accelerated development, reduced costs, and improved performance.

Unsupervised Defect Segmentation

The AI model learns from defect-free images, enabling it to identify defects as deviations from the norm. This eliminates the need for costly defect data collection or defect sample production. Defect images are only required for testing and verification, making it ideal for detecting anomalies when defects or defects categories are unknown in advance.

Instance Segmentation

Identify and separate individual objects within an image, assigning each a unique label. This method provides the precision needed for applications in object tracking, quality inspection, and robotic vision.

Supervised Defect Segmentation

Trains a model using both defected and non-defected images. The model learns to classify and segment various types of defects, distinguishing between different defect classes. By relying on labeled defect annotations, this model is ideal when precise defect identification is needed.

Object Detection

Object Detection can identify and label objects within an image, providing information about their positions, types, and quantity. This technique is useful for applications that need to recognize multiple objects at the same time.

OCR/Text Recognition

Extracts readable text from images, even deciphering deformed or skewed text. It transforms scanned documents, receipts, license plates, and other text-rich visuals into editable and searchable data.

Positioning

Accurately detects and output the location of objects within an image, providing precise spatial data. This capability helps automate processes such as robotic guidance and assembly line monitoring by enabling systems to understand the exact placement of parts and equipment.

Presence Detection

Identifies the presence of object in the image by searching for each individual item. Commonly used for checking for misplacement and missing parts.

Keypoint Detection

Identify and pinpoint distinctive features or locations in images. These keypoints serve as reference markers for robotic guidance.

Classification

Helps categorize objects within images into specific classes or labels and distinguish between defective and non-defective items.

Smart and fast

Train Your AI for Optimal Results

We understand the challenges involved in training AI models—it often requires significant resources, including large amounts of data and substantial human effort. DaoAI World is designed to help you overcome these obstacles with ease.

Unsupervised AI Learning With Few Positive Samples Alone

Unsupervised AI learning empowers AI to learn and extract valuable insights from just a few positive samples, without requiring negative or contrasting examples. By identifying patterns, structures, and relationships within the provided data, AI can generalize knowledge and make accurate predictions or classifications. Through this method, users can harness the power of AI to build smarter systems with less data preparation effort.

Data Requirement 
Learning Time
Unsupervised AI Learning
 1-20 positive examples alone With-in 1 minute
Edge AI
 Dozens of correct and incorrect image samples Several minutes
Typical Deep Learning
 100+ samples Several hours to days

Smart Labeling

By combining machine vision with intelligent algorithms, smart labeling allows users to annotate objects with a simple click - eliminating the need to manually trace edges by clicking and dragging.

video poster image
Comprehensive and flexible

Adaptability and Integration

DaoAI World simplifies the transition from your current platform to runtime deployment. We support direct upload of pre-annotated data (.json) files from existing projects and provide software development kits and multiple deployment method to accelerate implementation.

 
 

Transport Format Pre-labeled Data

Upload pre-annotated data to import existing labels and annotations, saving time and preventing duplicate work. 

 

 

 Support Format:

COCO JSON
Pascal VOC
YOLO v8 Pytroch
YOLO v5 Pytorch
VGG Image Annotator
JSON (VIA)

 
 

Flexible Integration

DaoAI offers versatile and efficient solutions to meet the varying needs of users with different computational and operational requirements.

 

  
Standard SDK:Supports C++, C#, and Python.
 
Self-Hosted Inference Server:Built upon the Standard SDK, this method enables setting up a local inference server that communicates via HTTP requests.
 
DaoAI World Hosted Inference Server:This method supports Python and utilizes HTTP API calls to perform inference using models hosted on the DaoAI World server, enabling simultaneous training and inspection.
(Only Available For DaoAI World Enterprise)
Seamless and real-time

Data Sharing and Monitoring

Supports real-time collaboration that enables team members to work on the same project. Allowing multiple users to annotate data, adding new training sessions or fixing errors, simultaneously. Bring consistency and efficiency in the annotation process.

Entire machine learning pipeline exists in one central repository for data monitoring. Shared accounts (“Workspaces”) house all source images, annotations, datasets and trained models (both ready to be deployed, and actively in-deployment).

Training Metrics
During training, it tracks mean average precision (mAP), precision, recall, and training loss. After training is done, track training time and single image inference time.
 
Process Tracking
Monitor dataset labeling and review completion.
 
Workflow Status
Provide an end-to-end workflow view, monitor pipeline stages from ingestion to deployment. 
 
Annotation Quality
Assess annotation accuracy and consistency.
Screenshot 2025-02-05 at 2.46.58 PM
build visual ai in 6 steps

From Training to Deployment

Screenshot 2025-02-14 at 2.03.37 PM
product line-up

Optimization and Scalability Tailored to Your Needs

ENTERPRISE

Beginner

ENTERPRISE

Regular

ENTERPRISE

Pro

INDUSTRIAL

Beginner

INDUSTRIAL

Regular

INDUSTRIAL

Pro

GPU

NVIDIA RTX 4080 16G x 4

NVIDIA RTX 4090D 24Gx 4

NVIDIA RTX 4090D 24Gx 8

NVIDIA RTX 4060 Ti

NVIDIA RTX 4080

NVIDIA RTX 4080 + RTX 4060 Ti

CPU

Intel Xeon Gold 6330 x2

Intel Xeon Gold 6330 x2

Intel Xeon Gold 6330 x2

Intel Core i7-10700K

Intel Core i7-10700K

Intel Core i7-10700K

Storage

480GB SATA SSD  2.5in x 1

3.84TB SATA SSD  2.5in x 5

480GB SATA SSD  2.5in x 1

3.84TB SATA SSD  2.5in x 5

480GB SATA SSD  2.5in x 1

3.84TB SATA SSD  2.5in x 5

1TB SSD M.2

1TB SSD M.2

2TB SSD M.2

RAM

32GB DDR4 x 8

32GB DDR4 x 8

32GB DDR4 x 8

16GB DDR4 x 2

16GB DDR4 x 2

32GB DDR4 x 2

APPLICATIONS

1
Keycap Defect Detection
Electronics

Unsupervised Defect Segmentation

Untitled design 2
Track Link Inspection 
Machinery Manufacturing

Instance Segmentation

3
Fruit Inspection and Sorting
Food & Agricultural

Object Detection

9
Precision Detection 
Robotic Automation

Keypoint Detection

5
Package Traceability
Packaging Industry

Instance Segmentation

7
Dental Braces Defect Detection
Pharmaceutical & Medical

Supervised Defect Segmentation

8
Wheel Stud Assembly Detection
Automotive Assembly Inspection

Presence Detection

4
Precise Positioning of Latch
Machinery Manufacturing

Positioning

6
Tablet Quality Inspection
Pharmaceutical & Medical

Classification

Learn More about

DaoAI World

View Data Sheet