The project is funded with almost 4 million euros by Horizon 2020, the European Union's largest innovation program to date.
TULIPP will focus on developing a reference platform for designers of vision-based systems. The project defines a set of recommendations for selecting relevant combinations of processing and communication resources, which should be instantiated on the platform to minimize energy consumption while reducing costs and time to market. 
The recommendations will address the inherent complexity of designing the next generation of embedded image processing applications emerging across a wide range of industries. From an application perspective, this complexity relates to the need to ensure high performance combined with increased energy efficiency within the context of embedded system design requirements. Regarding available hardware platforms, software developers must be able to easily handle issues of parallel programming of multicore devices and the heterogeneity of different programming models and APIs.
The reference platform description specifies what a hardware or software component must satisfy to comply with TULIPP. Within the framework of this project, we will use this description to develop an instance of the TULIPP reference platform, including a scalable, low-power board designed to meet typical size, weight, and power (SWAP) requirements for embedded systems, a real-time operating system (RTOS), low-power image processing libraries, and a toolchain that takes energy efficiency into account. Furthermore, TULIPP will develop three use cases as proof of concept and validation of the reference platform. These cases cover different industrial fields with complex image processing requirements: a medical X-ray imaging system for surgery, designed to significantly reduce radiation dose by 75%; an integrated intelligent automotive vision system for advanced driver assistance systems (ADAS) that, in addition to processing low-level images, intelligently interprets them to provide safer driving experiences; and an integrated image processing system for intelligent drones (UAVs) for search and rescue of survivors in the event of disasters. By the end of the project in 2018, TULIPP will have increased the peak performance per watt in image processing applications by a factor of 4 and the average performance per watt by a factor of 10. After the official end of the project, it is expected that in 2023 the increase could reach a factor of 100 for the peak and 200 for the average, respectively.
The members of the TULIPP consortium come from industry and academia. They are: Thales (France) as project leader and coordinator, along with Efficient Innovation SAS (France), Fraunhofer IOSB (Germany), Hipperos (Belgium), Norges Teknisk-Naturvitenskapelige Universitet (Norway), Ruhr-Universität Bochum (Germany), Sundance Multiprocessor Technology (UK), and Synective Labs (Sweden). Together, they provide the complementary knowledge and experience necessary to make the project a success.
TULIPP works closely with several standards bodies to propose the formal adoption, with a broad industry base, of new standards from its reference platform. TULIPP is also seeking to establish an Advisory Board with stakeholders in vision-based systems to review project outcomes and gradually help expand the range of image processing applications into other industry sectors. Companies interested in joining the Advisory Board can express their interest by sending an email to
"Image processing applications are extending across an ever-expanding range of industrial fields and reaching a level of complexity greater than ever before," said Philippe Millet of Thales and coordinator of the TULIPP project. "The TULIPP reference platform will lead to significant advances in systems integration, innovation processes, and energy management to address the challenges involved in increasingly complex machine vision systems.".
