Level up your computer vision workflows with a free hands-on workshop for your team! Book a workshop: hubs.ly/Q04nvSCk0 These hands-on workshops are delivered by Voxel51 computer vision experts. Both virtual and in-person formats. * 60 min virtual workshop * Half-day onsite workshop * Full-day onsite workshop and hackathon #mcp #skills #computervision #ai #artificialintelligence #machinevision #machinelearning #physicalai

#TestEngineering #softwarevalidation #machinevision

Join us on July 28 for the Paris AI, ML and Computer Vision Meetup! Pre-registration is mandatory as seats are limited - hubs.ly/Q04nvGgn0 Talks will include: * Finetuning VLMs for domain specific tasks - Amine Belhakimi at GoPro * Computer Vision at Nanoscale - Detecting, Segmenting and Analyzing Nanoparticles in microscopic images - Atif Anwer at (Ex) University Bourgogne Europe * Towards a Resolution- and Modality-Agnostic Transformers for Earth Observation - Guillaume Astruc at Imagine - ENPC * Building Real-World Computer Vision Systems with Voxel51 - Harpreet Sahota at Voxel51 * Efficient Image Generation through Smarter Data, Objectives, and Alignment - Lucas Degeorge at Ecole Polytechnique - Ecole des Ponts - AMIAD *********** Level up your computer vision workflows with a free hands-on workshop for your team! Book a workshop: hubs.ly/Q04nvLHd0 These hands-on workshops are delivered by Voxel51 computer vision experts. Both virtual and in-person formats. * 60 min virtual workshop * Half-day onsite workshop * Full-day onsite workshop and hackathon #mcp #skills #computervision #ai #artificialintelligence #machinevision #machinelearning #physicalai

🚀 𝐄𝐥𝐞𝐯𝐚𝐭𝐞 𝐒𝐦𝐚𝐫𝐭 𝐌𝐚𝐧𝐮𝐟𝐚𝐜𝐭𝐮𝐫𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐂𝐢𝐧𝐜𝐨𝐳𝐞'𝐬 𝐌𝐃-𝟑𝟎𝟎𝟎 𝐌𝐚𝐜𝐡𝐢𝐧𝐞 𝐕𝐢𝐬𝐢𝐨𝐧 𝐂𝐨𝐧𝐭𝐫𝐨𝐥𝐥𝐞𝐫 We're excited to share that @CtrlAutomation has featured Cincoze's MD-3000 Series, an iF Design Award-winning solution designed to tackle real-world machine vision challenges on the plant floor. Built for demanding industrial environments, the MD-3000 helps machine vision systems achieve higher performance, greater flexibility, and smarter edge AI deployment. 💡 𝐊𝐞𝐲 𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬: ✅ High Performance: 14th Gen Intel® Core™ processors and up to 96GB DDR5 memory ✅ Compact Design: Space-saving DIN-rail form factor at just 152 mm high ✅ Scalable Architecture: Flexible expansion for I/O, functional modules, storage, and application-specific configurations 👇Check out the full feature on Control.com to read more! control.com/news/cincoze-hig… #Cincoze #MD3000 #MachineVision #EdgeAI #SmartManufacturing #IndustrialAutomation #DINRailComputer

Join us on July 27 for the London AI, ML and Computer Vision Meetup.  Pre-registration is mandatory - hubs.ly/Q04nvCjv0 Talks will include: * UniLight: Unified Multi-Modal Lighting Representation - Zitian Zhang at Adobe Research London, Université Laval * Lessons from the Trenches of Agentic Engineering - John Adeojo at Brainqub3 * LoST: Level of Semantics Tokenization for 3D Shapes - Niladri Dutt at UCL | Adobe  * Material selection in 2D and beyond - methods, tricks and applications - Michael Fischer at Adobe * Building Real-World Computer Vision Systems with Voxel51 - Harpreet Sahota at Voxel51 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #opensource #london

Join us on July 23 for the monthly AI, ML, and Computer Vision Meetup! Register for the Zoom: hubs.ly/Q04nvRBH0 Talks will include: * Generative AI for Video Trailer Synthesis: From Extractive Heuristics to Autoregressive Creativity - Abhishek Dharmaratnakar at Google * Training-Free Object and Associated Effect Removal in Videos - Saksham Singh Kushwaha at University of Texas at Dallas * Making Agent Systems Observable, Reliable, and Testable - Adonai Vera at Voxel51 * Turning Models into Systems: AI Architecture That Works - Nikita Golovko at Siemens *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nvv5d0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Join us on July 22 for the Best of ICRA virtual Meetup! Register for the Zoom: hubs.ly/Q04nv5tr0 Talks will include: * Contrastive learning on 3d point clouds for geometric defect detection - Alexander Tarvo at University of Washington * A Semantic and Occlusion-Aware Gaussian Mixture Probability Hypothesis Density Filter - Jovan Menezes at Cornell University * An Annotation-to-Detection Framework for Autonomous and Robust Vine Trunk Localization in the Field by Mobile Agricultural Robots - Dimitrios Chatziparaschis at University of California, Riverside * vS-Graphs: Tightly Coupling Visual SLAM and 3D Scene Graphs Exploiting Hierarchical Scene Understanding - Ali Tourani at University of Luxembourg *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nv4mg0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

99.9% AUROC on MVTec sounds like a solved problem — until you deploy in a real warehouse and the numbers collapse. That's the gap Amazon's Kaputt dataset was built to expose. Read it here: hubs.ly/Q04nvHY90 * 230K images * 29K defective instances * 48K distinct objects Roughly 40x larger than MVTec, with the messy pose and appearance variation of actual retail logistics. The best state-of-the-art methods scored just 56.96% AUROC. Barely better than a coin flip. It powered the Retail Track of the VAND 4.0 Challenge at CVPR 2026, and our latest post walks through how to explore Kaputt with open source FiftyOne: grouped multimodal views, embedding-based exploration with CLIP UMAP, similarity search, and VLM captioning with FastVLM. #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Most vision models tell you how things moved after the fact. Forecasting motion — where an object goes before it gets there — is the harder problem, and the one robots and video models actually need. Ai2's PointMotionBench benchmarks exactly this. But it ships as raw .npz tracks and JSON — great for scoring, not great for understanding - hubs.ly/Q04nvDX90 So we loaded it into FiftyOne. Now you can: * Scrub ground-truth points frame by frame * Filter clips by object and caption * Spot tracker drift near reflections and occlusions * Auto-surface the model's worst predictions once you add them A benchmark you can't see is one you can't reason about. #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Join us on July 10 for day 3 of the “Best of CVPR” series. Register for the Zoom: hubs.ly/Q04nvSsZ0 Talks will include: * Advancing Generative Quality and Reasoning in Multimodal AI - Deepti Ghadiyaram at Boston University * HyperRealm: Hyperbolic Vision Language Models for Real-World Hierarchical Multimodal Understanding - Kathy Wu at Amazon * Cross-Modal Domain Adaptation using Semantic Parametric Mapping - Frank Bieder at FZI Forschungszentrum Informatik * WalkGPT: Pixel-Grounded Navigation Guidance for Pedestrians - Rafi Ibn Sultan at Wayne State University *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nvH--0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Level up your computer vision workflows with a free hands-on workshop for your team! Book a workshop: hubs.ly/Q04nvxjB0 These hands-on workshops are delivered by Voxel51 computer vision experts. Both virtual and in-person formats. * 60 min virtual workshop * Half-day onsite workshop * Full-day onsite workshop and hackathon #mcp #skills #computervision #ai #artificialintelligence #machinevision #machinelearning #physicalai

Join us on July 9 for day 2 of the “Best of CVPR” series. Register for the Zoom: hubs.ly/Q04nsTjg0 Talks will include: * Efficient Representation and Coding of Dynamic Light Fields - Joshitha Ravishanker at Indian Institute of Technology Madras * PHANTOM: Physics-Infused Video Generation via Joint Modeling of Visual and Latent Physical Dynamic - Ismini Lourentzou at University of Illinois Urbana-Champaign * LoST: Level of Semantics Tokenization for 3D Shapes - Niladri Dutt at UCL | Adobe * 3D Reconstruction Improves Weakly-Supervised Semantic Segmentation - Wolfgang Boettcher at Max Planck Institute for Informatics *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nsWzt0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Six cameras? A spinning LiDAR? A GPS trace? All capturing the same instant? You can build it: hubs.ly/Q04nvpCC0 We built a FiftyOne plugin inspired by a Motional and UCLA Henry Samueli School of Engineering and Applied Science animation for their nuReasoning dataset. The plugin enables you to play everything in lockstep: camera frame, bird's-eye LiDAR, and a live map of every agent, advancing together frame by frame. The fun part is watching a self-driving car think. The useful part is what it catches — a cuboid floating off its object, an agent on the map that isn't where the camera says it is. Calibration bugs you'd otherwise hunt for in spreadsheets jump right out. Built on grouped nuScenes data with the FiftyOne plugin framework. Works for robotics, drones, anything with a moving platform and a pile of sensors. #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Join us on July 8 for day one of the “Best of CVPR” series Register for the Zoom: hubs.ly/Q04nsHlz0 Talks will include: * CylinderDepth: Cylindrical Spatial Attention for Multi-View Consistent Self-Supervised Surround Depth Estimation - Samer Abualhanud at Leibniz University Hannover * Your ViT is Secretly Also a Video Segmentation Model - Daan de Geus at Eindhoven University of Technology * LinkedOut: Linking World Knowledge Out of Video LLMs for Next-Generation Video Recommendation - Haichao Zhang at Northeastern University * Some Modalities Are More Equal Than Others: Understanding and Improving Multimodal Integration in MLLMs - Tianle Chen at Boston University *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nsNlQ0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Join us on July 8 for day one of the “Best of CVPR” series Register for the Zoom: hubs.ly/Q04nvr3k0 Talks will include: * CylinderDepth: Cylindrical Spatial Attention for Multi-View Consistent Self-Supervised Surround Depth Estimation - Samer Abualhanud at Leibniz University Hannover * Your ViT is Secretly Also a Video Segmentation Model - Daan de Geus at Eindhoven University of Technology * LinkedOut: Linking World Knowledge Out of Video LLMs for Next-Generation Video Recommendation - Haichao Zhang at Northeastern University * Some Modalities Are More Equal Than Others: Understanding and Improving Multimodal Integration in MLLMs - Tianle Chen at Boston University *********** Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place. “pip install fiftyone” is all it takes to get started - hubs.ly/Q04nvvSZ0 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

A 768-dimensional vector doesn't mean much to the human eye. Now you can actually see it - hubs.ly/Q04nqHk70 Ai2 OlmoEarth v1.2 just made satellite foundation models way cheaper to run — 3× less training compute, 2.9× fewer ops at inference, zero accuracy loss across 13 benchmarks. Pair it with Voxel51's open source FiftyOne library and the story gets better: drop in the embeddings, reduce with UMAP, and watch tiles self-organize into clusters with no labels at all. Filter to a country on the map and the image grid embedding scatterplot snap into sync instantly — geography, pixels, and embedding space, linked live. Cheaper models tools to actually see what they're producing = planetary-scale insight on a laptop. Fully open: weights, training code, dataset, notebook. #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents

Level up your computer vision workflows with a free hands-on workshop for your team! Book a workshop: hubs.ly/Q04nsVmF0 These hands-on workshops are delivered by Voxel51 computer vision experts. Both virtual and in-person formats. * 60 min virtual workshop * Half-day onsite workshop * Full-day onsite workshop and hackathon #mcp #skills #computervision #ai #artificialintelligence #machinevision #machinelearning #physicalai

Join us on July 28 for the Paris AI, ML and Computer Vision Meetup! Pre-registration is mandatory as seats are limited - hubs.ly/Q04nsDkw0 Talks will include: * Finetuning VLMs for domain specific tasks - Amine Belhakimi at GoPro * Computer Vision at Nanoscale - Detecting, Segmenting and Analyzing Nanoparticles in microscopic images - Atif Anwer at (Ex) University Bourgogne Europe * Towards a Resolution- and Modality-Agnostic Transformers for Earth Observation - Guillaume Astruc at Imagine - ENPC * Building Real-World Computer Vision Systems with Voxel51 - Harpreet Sahota at Voxel51 * Efficient Image Generation through Smarter Data, Objectives, and Alignment - Lucas Degeorge at Ecole Polytechnique - Ecole des Ponts - AMIAD *********** Level up your computer vision workflows with a free hands-on workshop for your team! Book a workshop: hubs.ly/Q04ntbw60 These hands-on workshops are delivered by Voxel51 computer vision experts. Both virtual and in-person formats. * 60 min virtual workshop * Half-day onsite workshop * Full-day onsite workshop and hackathon #mcp #skills #computervision #ai #artificialintelligence #machinevision #machinelearning #physicalai

Join us on July 27 for the London AI, ML and Computer Vision Meetup.  Pre-registration is mandatory - hubs.ly/Q04nqtps0 Talks will include: * UniLight: Unified Multi-Modal Lighting Representation - Zitian Zhang at Adobe Research London, Université Laval * Lessons from the Trenches of Agentic Engineering - John Adeojo at Brainqub3 * LoST: Level of Semantics Tokenization for 3D Shapes - Niladri Dutt at UCL | Adobe  * Material selection in 2D and beyond - methods, tricks and applications - Michael Fischer at Adobe * Building Real-World Computer Vision Systems with Voxel51 - Harpreet Sahota at Voxel51 #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #opensource #london

Most self-driving datasets tell you what's in the scene. nuReasoning tells you what the car decided — and what it ruled out - hubs.ly/Q04nsNJc0 From Motional and UCLA Henry Samueli School of Engineering and Applied Science: ~20,000 real-world long-tail driving clips, annotated with the alternatives the car considered, each labeled safe, suboptimal, or unsafe, with the reason it was chosen or rejected. That counterfactual layer is the whole point. With Voxel51's open source FiftyOne app, that reasoning becomes something you can watch. Load a clip as a grouped video dataset, play all eight cameras and the BEV map in sync, scrub to the decision frame — and see the unsafe option the car rejected, color-coded by risk, with the why. Not a benchmark number. An interpretable window into how a car thinks on exactly the edge cases that matter for safety. #computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #physicalai #mcp #agents