Ficha, a prominent Japanese AI development company specializing in automotive and image recognition technologies, has allegedly been compromised, resulting in the leak of its corporate GitHub repositories. The breach reportedly exposes the company’s core technical framework alongside highly sensitive, customer-specific project files.
The allegedly compromised data encompasses a massive collection of the company’s foundational technology and client collaborations. According to the leak, the exposed information includes:
-
Proprietary SDK Core Platforms (sdk2, SDK4, and AIFramework)
-
Advanced Driver Assistance Systems (ADAS) core algorithms (including FichaDet and LaneDet)
-
Driver Monitoring Systems (DMS) algorithms and neural networks
-
Customer-specific project code for major global brands, including Bosch, Clarion, JVC, and Toyota
-
Core Optical Character Recognition (OCR) projects and LLM integration tools
-
Hardware platform adaptations for systems like NVIDIA Jetson and Renesas V4H
