A research team from Bar-Ilan’s Department of Computer Science, led by Dr. Dvir Samuel and Prof. Gal Chechik (also of NVIDIA), has developed OmnimatteZero, a new method for separating objects from video backgrounds without the need for heavy training or optimization.

Presented at SIGGRAPH Asia, the technology preserves complex visual details like fur, reflections, smoke, and water ripples while avoiding the huge datasets and computing power typically required. Instead, it uses advanced image-completion techniques with temporal-spatial tracking to maintain background consistency.

The system enables “visual composting,” allowing elements like a swan with its reflection to be seamlessly moved to another scene, or a background reused naturally. Unlike current methods, OmnimatteZero runs quickly and efficiently using existing video generation models, making it more practical for editors, content creators, advertisers, and everyday users.

The brain, one of the most sophisticated and complex organs, remains a mystery in many aspects of its function. Scientists in Switzerland and China developed a groundbreaking hybrid imaging platform—HyFMRI—to simultaneously capture the dynamic activity of neurons, astrocytes, and hemodynamic responses. The technology, demonstrated in a mouse model, enables unprecedented studies of brain activity, opening new frontiers for understanding neuro-glial-vascular coupling in health and disease.

This work proposes a novel lower-limb motion capture system that, for the first time, combines a flexible pressure sensor array with a Transformer-based temporal regression model. The system enables accurate estimation of lower-limb joint positions using only insole-embedded sensors.

Flexible fiber sensors, with their excellent wearability and biocompatibility, are essential components of flexible electronics. However, traditional methods face challenges in fabricating low-cost, large-scale fiber sensors. In recent years, the thermal drawing process has rapidly advanced, offering a novel approach to flexible fiber sensors. Through the preform-to-fiber manufacturing technique, a variety of fiber sensors with complex functionalities spanning from the nanoscale to kilometer scale can be automated in a short time. Examples include temperature, acoustic, mechanical, chemical, biological, optoelectronic, and multifunctional sensors, which operate on diverse sensing principles such as resistance, capacitance, piezoelectricity, triboelectricity, photoelectricity, and thermoelectricity. This review outlines the principles of the thermal drawing process and provides a detailed overview of the latest advancements in various thermally drawn fiber sensors. Finally, the future developments of thermally drawn fiber sensors are discussed.

A research team from the South China University of Technology has developed an innovative statistical modeling approach that accelerates the development of advanced rare-earth-doped laser glasses. Applying neighboring glassy compounds (NGCs) model, the team accurately predicted the local structural environments and luminescence properties of complex glass systems, reducing experimental trial-and-error. The NGCs model was used to establish the composition-structure relationship and populate the composition-property space. Finally, multi-luminescence property charts are generated to select compositions that satisfy multiple constraints, thus facilitating the rational design of chemically complex laser glasses for targeted applications. This versatile methodology paves the way for discovering next-generation laser materials with superior performance, expanding the horizons of glass science and technology.

A research paper by scientists at The University of New South Wales presented a new hydraulic-driven dual soft robotic system featuring a 3 DOF-soft cutting arm (SCA) and a 3-jaw teleoperated soft grasper system (TSGS).

The research paper was published on Jun. 12, 2025 in the journal Cyborg and Bionic Systems.