Beamr and Intempora to demo lossless compression for AV data

Beamr Imaging Ltd. (Nasdaq: BMR) and Intempora, a dSPACE company, announced they will demonstrate Beamr's ML-Safe lossless compression integrated into RTMaps, a real-time multisensor middleware used in autonomous vehicle and advanced driver-assistance systems development.

The demonstration is scheduled for AutoSens Europe in Barcelona, September 22–24, and the dSPACE User Conference in Santa Clara, California, on September 23.

According to the companies, Beamr's lossless compression supports 12-bit Bayer format — the raw output of test vehicle cameras — running in real time on the GPU inside a vehicle's data logger. The press release states that previous encoders in test vehicles did not support the 12-bit Bayer format, leaving teams to store files as uncompressed raw data or discard them.

Beamr's internal testing showed a 47% file size reduction on recordings from eight 8-megapixel cameras. The company says this effectively doubles capture capacity per vehicle and reduces the time needed to transfer recordings to the cloud.

RTMaps users will have access to two compression modes: lossless, which preserves data bit-for-bit, and content-adaptive, which the company says reduces video data by up to 50% while maintaining machine learning model accuracy.

"The volume of sensor and video data our users move through RTMaps, and then store and manage in IVS, keeps growing," said Nicolas Du Lac, CEO of Intempora. "Compression is essential to keep that data workable, and together with Beamr we are demonstrating it in two modes — lossy when a smaller file with preserved ML accuracy aligns with their needs, and lossless when the data has to stay bit-exact."

Peter Ngure, Product Manager at dSPACE, said efficient compression at the point of capture "without compromising data integrity, is an important building block," adding that the collaboration gives RTMaps users more options to manage data across the toolchain.

Haggai Barel, Beamr's COO, said teams can apply lossy content-adaptive compression to reduce data in the cloud and for simulation and training datasets, and can now also use lossless compression at the data logger stage.

You May Also Be Interested In

• AEVEX says Navy GARC orders were not halted, contract now at $350M
• IBM stock gains after earnings beat expectations
• Santacruz Silver posts Q3 2026 output of 1.5M silver ounces