Live Stream Date/Time: Thursday, October 1st, 2026 @ 9am Pacific. Akshat Mandloi, co-founder of Smallest.ai, joins OpenCV Live to answer a question most of us have wondered on hold with customer service: why can you still tell it’s a bot? His answer isn’t “the model’s too small.” After three generations of voice AI, less than…
Click-Ins runs inspections for insurers, rental fleets and dealerships from ordinary phone photos. The models behind it are trained almost entirely on rendered data: 3D vehicle models with procedurally generated dents, scratches and deformations across makes, colors, lighting and backgrounds, auto-annotated at the pixel level. In this episode CEO Dmitry Geyzersky will cover how the…
OpenCV’s DNN module lets you run a trained neural network inside the same application that reads images, processes video, and draws results. You can export a model to ONNX, load it with OpenCV, and use its predictions without loading the training framework in your application. OpenCV 5 expands that workflow with a new inference engine…
Planned work will focus on native builds, sustainable build-and-test infrastructure, and performance optimization for the open-source developer community PALO ALTO, Calif., Sept. 14, 2026 — The Open Source Vision Foundation (OSVF), the nonprofit organization that operates OpenCV, today announced plans to enhance OpenCV support for Windows on Snapdragon®. As part of this effort, OSVF is…
Pengyuan Li of IBM Research returns to OpenCV Live, following his August 2025 appearance where he introduced Granite Vision, IBM’s lightweight open vision-language model built for enterprise document understanding. This time the focus shifts from the model to the workflow: how Granite Vision pairs with Docling, IBM’s open-source document conversion toolkit, to turn PDFs, scans,…
The goal behind this session is to explain to developers how Visual SLAM maps differ from VPS maps, and how to approach challenges such as short-term and long-term relocalization. We will also explore the differences between machine-readable and human-readable maps, and why both need to coexist to support orchestration and coordination tasks occurring within the…
Camera calibration is a familiar pain to anyone who works with computer vision. It can be hard to know where to start, and even difficult to know when you’re “done.” This week we welcome back Luxonis to the show, and Cenek Albl (VP of Computer Vision) will show us how they handle the most dangerous…
The 3D-Object Perception Transformer (3PT) unifies detection, segmentation, and 6DoF pose estimation into two multi-view, RGB-only transformers. Demonstrating exceptional accuracy and cross-domain robustness, it placed first by significant margins in both the Industrial Robotics and AR/VR tracks of the BOP 2025 challenge at ICCV. Today, the 3PT architecture is actively deployed in real-world industrial robotic…
This week join Jason Ren of Ai2 (Allen Institute for AI) for MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction, a live look at how open multimodal models move past describing a scene to predicting where things go next. Building on the Molmo family’s pointing and tracking capabilities, MolmoMotion takes a plain-language instruction and…
Global computer vision challenge opens to developers, researchers, and teams worldwide with prizes totaling $10,000 [Palo Alto, CA] — The Open Source Vision Foundation (OpenCV), home to the world’s most widely used open-source computer vision library, today announced the OpenCV AI Competition 2026 in collaboration with Amazon Web Services (AWS). It is the latest in…