Camthink has announced the NeoEyes NE503, a $1199 4K Edge AI camera built around the Hailo-15H SoC featuring a quad-core Cortex-A53 CPU and 20 TOPS AI accelerator. It pairs a Sony IMX678 4K sensor with an 8–32mm F1.6 auto-zoom lens, 8GB LPDDR4, and 64GB eMMC. The camera runs embedded Linux via Yocto, supports containerized apps, and offers a REST API, Python SDK, and CLI tools. Pre-installed models include YOLOv8n for person detection, face landmarks, and CLIP Image Encoder. It supports PoE 802.3AT, is IP67-rated, and operates from -40 to 60°C. Pre-orders are open with shipping expected in late July 2026.
Nguồn: https://www.cnx-software.com/2026/06/25/neoeyes-ne503-a-1199-edge-ai-camera-based-on-hailo-15h-20-tops-soc. 8sync News chỉ tóm tắt và dẫn link; bản quyền nội dung thuộc tác giả và nguồn gốc.

Netflix giới thiệu hai mô hình chỉnh sửa video AI giai đoạn đầu là Vera và VOID. Vera sử dụng mô hình diffusion phân lớp, chỉ tái tạo vùng chỉnh sửa (kèm alpha matte) thay vì toàn bộ clip, bảo toàn nội dung chưa chỉnh sửa. VOID chuyên xóa vật thể trong video với kỹ thuật inpainting hợp lý vật lý, tái tạo cảnh thực tế khi vật thể bị loại bỏ. Cả hai mô hình đều vượt trội so với các phương pháp hiện có trong nghiên cứu.
Lập trình viên muốn phát triển các giải pháp AI tiên tiến trong xử lý video nên tham khảo để hiểu cách thiết kế mô hình hiệu quả như Vera và VOID, từ kiến trúc đặc biệt đến kỹ thuật điều khiển chi tiết để nâng cao chất lượng và tính khả thi của các ứng dụng AI video trong tương lai.
MYiR has launched the MAC-B5760, a fanless Edge AI industrial box PC built around the Rockchip RK3576 SoC featuring a 6 TOPS NPU. It ships with 8GB LPDDR5 and 64GB eMMC by default, offering dual Gigabit Ethernet, four USB 3.0 ports, HDMI 2.1, Mini DisplayPort, optional WiFi 6/BT 5.4, and expansion via M.2 and mini PCIe slots. An optional Rockchip RK1828 M.2 AI accelerator module (20 TOPS) supports LLM/VLM inference for models up to 7 billion parameters. Software support includes a Linux 6.1 BSP with Debian, Yocto, and Yocto Preempt RT images. The standard model starts at $299, while the variant with the RK1828 AI module is $599.
Researchers at EPFL's NeuroAI Lab have developed AI-based topographic neural network models that predict optimal brain stimulation patterns to evoke perception of complex visual objects — such as faces and houses — rather than just simple light flashes. The models were validated in live trials on sighted monkeys in Amsterdam, showing that model-guided cortical stimulation can bias visual object perception in predictable ways. While the team cannot yet create object perception from scratch (without any visual input), this is the stated next goal. The approach could also extend to improving cochlear implants for auditory prosthetics.
Berlin startup Almetra (formerly Deltia) has raised €16.3M in Series A funding to expand its AI-powered factory floor analytics platform. The company mounts cameras above assembly lines at manufacturers like Bosch, Siemens Energy, and ABB, converting video footage into live production data — cycle times, output rates, equipment utilisation — without requiring IT system integration. Customers report productivity gains of 15–19%. The round was led by blisce/, with participation from Merantix Capital and others. Almetra has been accepted into Google DeepMind's Robotics Accelerator and an AWS/Nvidia/MassRobotics Physical AI Fellowship, positioning it as a potential data source for industrial robotics. The company plans to use the funding to expand into the US and build out robotics applications.
Graperain has announced the GR1126MB development board, built around a 42x42mm GR1126B stamp-hole system-on-module powered by the Rockchip RV1126B SoC. The SoC features a quad-core Cortex-A53 CPU, a 3 TOPS NPU supporting TensorFlow, ONNX, PyTorch, and Caffe, plus a 12MP ISP and AI-ISP. The module offers up to 4GB LPDDR4x RAM and 256GB eMMC, with extensive I/O including dual MIPI CSI, Gigabit Ethernet, USB 3.0, CAN Bus, and more. The development board adds WiFi/BT, optional 4G LTE, a 40-pin GPIO header, and dual camera connectors. Software support includes BuildRoot and Debian 12, along with an AI framework for object detection, face recognition, and edge analytics. No pricing has been disclosed.
BEVPoolV3 is a new CUDA kernel optimization for bird's-eye-view (BEV) pooling used in autonomous vehicles and robotics. The post walks through a practical GPU optimization workflow: classify whether the working set fits in L2 cache, remove redundant scatter traffic via a five-array INT32 scatter map, implement interval-owned scatter-reduce to avoid atomics, and validate with NVIDIA Nsight Compute. On RTX PRO 6000 Blackwell Max-Q (large L2), BEVPoolV3 FP8 achieves up to 42x speedup over the V2 baseline. On RTX A6000 (small L2, DRAM-bound), the adapted FP16 path reaches 19x speedup. The post also explains why FP8 outperforms NVFP4 for L2-resident scatter-reduce workloads, and how the same methodology applies to sparse embeddings, voxelization, and other irregular memory-bound kernels.
MIT Sports Lab, co-founded in 2015, has become a key technology partner for major sports organizations. The lab played a central role in validating FIFA's semi-automated offside technology (SAOT) used at the 2022 World Cup, processing over 108,900 skeletal data points per second to ensure accuracy. Beyond soccer, the lab developed an Expected Action Value (EAV) metric for the NBA to quantify player decision-making quality, helped Adidas optimize 3D-printed midsole designs using biomechanical models, and conducted a COVID-19 stadium attendance analysis for the NFL. The lab bridges academic research and industry needs, connecting MIT students and faculty with professional sports organizations.