Artificial Intelligence
Case Studies
- Project Name: Retail Traffic Counter
- customer: Japanese Company
- Description: Counting people entering and exiting a store help boost in-store analytics & facilitate marketing segmentation. This is a problem of the detection and tracking people from surveillance videos. In order to solve the problem of detecting people in each video frame, Deep Learning approach is used. In details, a detection engine is built by making uses of TensorFlow’s Object detection API/ Faster R-CNN. After recognizing people and PeopleID is generated, SORT/ deep SORT, a tracking algorithm for 2D multiple object tracking in video sequences, is applied for real-time tracking people. The project is now going to evaluation phase.
- Technical Used: Python, OpenCV 1111, Scikit-Learn222022, TensorFlow弊社では次の業務にて豊富な経験を持っております。, Yolov3222
- Project Name: AI-OCR – Printed text and Handwriting Recognition
- customer: Japanese Company
- Description: Develop an offline handwriting recognition engine that can automatically read handwritten prescriptions in English from scanned or captured images. The perfect combination of a deep learning neural network and a deep learning neural network will make it possible to process the content of documents with even greater accuracy and clarity. Text can be extracted with high accuracy regardless of the quality of the original document, whether printed text, handwriting, or low-quality images.
- Technical Used: Python, Tensorflow, OpenCV
- Project Name: Rubber Stamp Removal
- customer: Japanese Company
- Description: The project aims to develop an engine that can automatically detect & remove rubber stamp from scanned/captured document images.. There are many challenging that we have to cope with in this project. For example, no any standards for rubbers so far (e.g. the variety of rubber shapes, colors) or especially dealing with both scanned and captured images are also a big challenging. After trying several methods and considering between two important metrics (accuracy and performance), we finally deployed YOLOv3 for object detection step and making use of K-means scikit-learn and OpenCV for output generation.
- Technical Used: Python, OpenCV, Yolov3, Scikit-learn
- Project Name: Reading Insurance Card
- customer: Japanese Company
- Description: Reading information from Insurance Cards is a problem of OCR (Optical Character Recognition) for Japanese Words. In order to read the 3 information from Insurance Card, Deep Learning approach is used. In details, we use Tesseract engine to read all information from the card and then employ some combination/ modification features to improve the recognition results. The improvement is basically based on “try and improve” process based on real data that we collected from Internet. • Categorial information (for example, MNIST: 0-9) should follow a categorical distribution: 𝑐1∼𝐶𝑎𝑡 (𝐾=10, 𝑝=0.1) • Shape information (rotation, width) should follow a uniform distribution, for example: 𝑐2, 𝑐3∼𝑈𝑛𝑖𝑓 (−1,1)
- Technical Used: Python, OpenCV, Yolov3, Scikit-learn
- Project Name: Dental Classification
- customer: Japanese Company
- Description: The project aims to develop an engine that can automatically detect & remove rubber stamp from scanned/captured document images.. There are many challenging that we have to cope with in this project. For example, no any standards for rubbers so far (e.g. the variety of rubber shapes, colors) or especially dealing with both scanned and captured images are also a big challenging. After trying several methods and considering between two important metrics (accuracy and performance), we finally deployed YOLOv3 for object detection step and making use of K-means scikit-learn and OpenCV for output generation.
- Technical Used: Python, Tensorflow, OpenCV, Yolov3
- Project Name: Pneumonia Detection
- customer: Japanese Company
- Description: The project aims to build an algorithm to detect a visual signal for pneumonia in medical images. Specifically, the algorithm needs to automatically locate lung opacities on chest radiographs. The dataset size is set around 23,124 images while the validation size is 2,560 images. To solve the problem, we built our own U-Net with the enhancement of resblock to improve the accuracy of the algorithm. The result based testing dataset (1000 images) is very positive (f2 score ~0.2)
- Technical Used: OpenCV, U-net, Keras, OpenCV
- Project Name: Big Data Platform & Video Recommendation System
- customer: Viet Nam Company
- Description: Big Data Platform: Craw, collect, store, and analyse data from a variety of different data sources in real-time. Provide wide range of “bigdata reports” in real-time for the Company key stakeholders in order to help them make quick and accurate decisions to improve the business outcome. Video Recommendation System: The Recommendation System for Video is developed upon Bigdata platform/technologies. This system gives personalised recommendation for customers on Web, OTT App, and Set-top-box platforms based on the customer history, the information of genre, content, and other similar user history also giving upcoming trending for customer.
- Technical Used: Hadoop, Spark, Cassandra, Kafka, HBase
Our Service Offerings
Our priority is to satisfy clients with high quality services
Artificial Intelligence
We believe that AI will still be the breakthrough technology of the new era. Thus, Heligate has been strongly investing to strengthen our AI development capability to meet your requirement.
We particularly have much experience on multiple subjects in AI-OCR/AI-OMR, while we also be able to deal with overall domains: Computer Vision・NLP・Recommendation System・Voice Recognition (For Ex: Spell Checking）・Text Classification・Image Recognition & OCR in Healthcare…
Web & Mobile App Development
We have an abundant resource with experienced web developers (3-5 years x 50%) with various skills.
Among those, according to the demand of the market, we are now focusing to build up a larger resource of Angular, ReactJS, VueJS, NodeJS and .Net.
Our capable development domain widely varies at B2B＆B2C E-commerce site, Stock Warehouse Management System, Logistics, HR Management System, Matching site, E-learning site, Reservation System, ect… with the deep understanding of Japanese Payment System such as: GMO Payment, PayPay…
To each project requirement, we will study and submit you the most cost-effective with the best performance solution by utilizing our experience and strong skill, and flexibly adapt open CMS services; libraries; SQL Server, PostgreSQL or MongoDB for DB structure.
As the part of the whole development service, we have experienced on System Migration Projects, such as VB6 to VB.NET or Flex to Angular. And we also receive any requirement if you just need to order us the Prototype Creation (Prototyping Service).
In accordance with Web development, we also invest the abundant resource to provide Mobile development service.
Our mobile team can handle both Native App and Cross-platform App.
For Native App, we can handle with both Objective-C & Swift for iOS and Kotlin/C#/Java for Android.
For Cross-platform, we can handle both by Flutter and React Native. Meanwhile, we propose you to apply Flutter because of its strong points rather than React Native.
Our capable development domains are widely varied with the same web system’s domains.
To meet the demands, we always try to update with the latest technology to offer, such as: AR/VR, Text-to-speech, Speech-to-text, ect…
_1000.png)
Data Annotation
Being aware of the importance of Data Processing in any AI subjects, Heligate started providing the Data Annotation Service with the service brand name “Helinotation” utilizing our self-promoted automatic labelling tool, from 2021 May.
The input data can be in various formats such as txt., pdf, static video format, motion video format… and the output can be adjusted into formats according to your request.
Using our tool, you can interact directly with labelled data released real time to check the progress and quality, and you can also use it partly to train the model at the same time.
Heligate owns the data labelling team who can understand Japanese well, which is helpful when the project needs the deep understanding of Japanese daily life culture or any other specific business domains.
_1000.png)
After a half of the year from service launching, a number of Japanese customers have used the service with 100% satisfaction of our price, quality and duration.
Maintenance
Along with the development service, upon understanding the technical stacks and the business flow of the system, we offer the Maintenance Service for the Software to the customers.
To the customers who newly order the Maintenance Service, we will start by doing the system and business research before proposing which Maintenance scopes should be, together with your defined new functions to be developed.
Normally the main scopes of this service are:
- Defect Diagnosis and Treatment
- Software monitoring & trouble shooting
- Helpdesk about the Software System
- Development and Deployment of additional functions
Based on these above scopes, we will offer you the most appropriate maintenance team and price, with the common contract type is Time & Material Contract (so called: quasi-mandate contract)
In terms of scopes related to Server side, we cooperate with a local strategy partner company, who is the agent of Amazon Web Service (AWS) in Vietnam, but can also provide the service for Google Cloud Platform (GCP).
Labor Dispatch
We cooperate with a member company in Heligate Group to provide “Labor Dispatch Service” with 02 main markets which are Japan and Taiwan.
Our customer domain varies in many domains such as: Mechanical, Machinery Repair, Food Processing, Agricultural Works and Nursing.
The target of dispatch is in 03 types:
- Specific Profession
- Intern
- Study Abroad
With the professional process built by the rich-experienced team who has been working with many customers in Japan and Taiwan, we commit to provide the quality service by the capability of:
- The abundant labor resource which can adapt to variety of domain.
(*) Our network and labor database is utilized by “Labor Dispatch Matching Service Platform”, a platform integrated with AI Bot to use on PC, Mobile devices, which was 100% self-developed by our in-house developers.
- Our standard “In-house Training Process”: the labor is pre-trained to master the basic language skill, as well as the understanding of the culture and behavior of the destination country.
- Guarantee the dispatch schedule with fulfilled legal documents.
- Very competitive price and high responsibility.
In the near future, in addition to the dispatching domains above, we will consider to build up a system to dispatch the IT Engineers from Vietnam, which will help a lot to the shortage of IT resource in Japan.
Investment Promotion Support
Heligate cooperate with our strategy partner company to provide the “Investment Promotion Support Service” mainly to Japanese Enterprises who are planning to set up their Company, Representative Office, Joint Venture or Factory in the north of Vietnam.
The detail of the service includes:
- Investment Support: Consulting and Handling Procedures for
+ Establishment of Legal Corporation, Branch, Joint venture, Representative Office in Vietnam
+ Setting up Factory procedures and other related procedures
+ Registration for Tax, Social Insurance; support service in terms of Tax, Accounting and Social Insurance execution.
- Trade and Marketing Promotion Support:
+ Support Market Researching, creating network and source finding for both import-export between Japan & Vietnam.
+ Business support as a connecting bridge between Japanese and Vietnamese companies (for example: Japanese Translation, Legal check, customs clearance, logistics…)
With our specialist team, who has much experience working with Japanese customers, we can offer a wide range of support with high quality and flexibility.
FIND MORE ABOUT US
top