Achievements in the development of AI, 3D analysis, and local LLM accumulated since 2018.
This is a document introducing the AI development achievements of ANGEWORK Inc. It includes examples of projects undertaken in fields such as healthcare, manufacturing, agriculture, and research, based on AI and 3D analysis technologies cultivated since 2018. The document features case studies on intraoral 3D data segmentation, position and rotation estimation from intraoperative images, plant growth assessment, point cloud and photogrammetry analysis, integration of astrology engines with LLM, and local AI. Additionally, it includes a verification case where Gemma 4 E2B was integrated into the iPhone 16 Pro, enabling Japanese conversation without external communication during inference. You can see our development policy, which enhances accuracy, explainability, and practicality by not solely relying on AI for judgment but combining it with definitive logic and specialized techniques.
Inquire About This Product
basic information
Fields of expertise: AI and machine learning, image recognition, 3D analysis, point cloud and photogrammetry, VLM and LLM, local AI, on-device AI. Supported environments: Windows, iOS, Web, cloud, on-premises. Development technologies: Python, Unity, Swift/SwiftUI, various AI frameworks, etc. We utilize image, video, sensor data, 3D data, and specialized business data to support everything from PoC to integration into existing systems and the development of business applications. Depending on the application and accuracy requirements, we design by combining AI with deterministic logic and dedicated engines.
Price information
Pricing will be individually estimated based on the development content, target data, required accuracy, compatible environments, and the extent of integration with existing systems. We can accommodate everything from small-scale technical verifications (PoC) to full-scale development and AI integration into existing systems. Please consult us first regarding your challenges and budget.
Delivery Time
※The delivery date varies depending on the development content and the scope of verification. We will confirm your requests and propose an individual schedule from PoC to full-scale development.
Applications/Examples of results
Focusing on the fields of healthcare, manufacturing, agriculture, and research, it can be utilized for area determination of intraoral 3D data, position and rotation estimation from intraoperative images, plant growth assessment, point cloud and photogrammetry analysis, and more. Additionally, it supports text generation by combining specialized engines with LLM, local AI that does not transmit confidential information externally, and on-device AI that can be used in environments without communication. We provide support from technical verification through PoC to AI integration into existing systems and business applications, as well as full-scale development.
Detailed information
-
We introduce ANGEWORK's AI development policy, which combines AI's flexible recognition and text generation capabilities with definitive logic based on formulas and judgment rules. Rather than solely relying on AI for decision-making, we appropriately design image analysis, 3D analysis, dedicated engines, and business rules according to the application and required accuracy. We develop AI systems that can be utilized in the field, focusing on the medical, manufacturing, agriculture, and research sectors, equipped with explanatory power and reproducibility.
-
This is a technical example of analyzing plant height, trunk and root thickness, branching, and shape changes from images and 3D data obtained using smartphones and other devices. It aims to reduce visual confirmation and recording tasks performed by humans and to establish a system for continuously and quantitatively understanding growth conditions. It addresses everything from designing shooting methods to data analysis and the construction of judgment systems for applications in seedling cultivation, agriculture, green space management, and research.
-
We analyze point cloud and photogrammetry data obtained from LiDAR, images, videos, etc., to determine the shape, position, rotation, and state changes of the target object. By combining our proprietary 3D analysis technology with AI, we can expand into fields such as healthcare, manufacturing, equipment maintenance, robotics, agriculture, and spatial monitoring. We provide consistent support from the consideration of acquisition methods to 3D data generation, AI judgment, and integration with existing systems.
-
We have developed an app that incorporates Google's lightweight LLM "Gemma 4 E2B" into the iPhone 16 Pro, allowing for conversations in Japanese without external communication during inference. This includes not only the integration of the model but also the conversation interface using SwiftUI, history saving, and stabilization of operation on actual devices. We envision applications in environments with communication restrictions and confidentiality requirements, such as hospitals, research facilities, factories, underground areas, mountainous regions, and during disasters.
-
The astrology engine "MoonFace" logically calculates celestial positions and astrological conditions, while the LLM is responsible for explaining and articulating the results. By not entrusting specialized calculations and judgments to generative AI and instead dividing tasks according to each's strengths, it helps to reduce oversights and fluctuations in judgment. Astrologers can utilize accurate calculation results to quickly create their own unique readings and advice.
catalog（1）
Download All Catalogs
Company information
Our company has received numerous contract development projects from major corporations in the fields of software development and research utilizing advanced technologies. We have experience with WebGL, Unity, LiDAR, medical smart glasses/watches, and drones. Recently, there has been an increase in projects related to machine learning (AI) and ChatGPT (API integration, prompts). We also have multiple achievements in R&D for high-difficulty projects and special cases. Starting in fiscal year 2024, we have incorporated AI support into our development environment, successfully achieving efficiency and cost reduction.
