AI News
CV Model
Jul 13, 2026
SenseTime open-sourced SenseNova-Vision, a unified model for detection, segmentation, depth, structured understanding, and multiview 3D tasks, plus a 50-million-sample visual instruction corpus.
Robotics
Jul 16, 2026
Sunday Robotics reported that its ACT-2 model let the Memo home robot fold unseen clothes in unfamiliar homes with 99.1% success, transferring skills from human demonstrations without new setup.
Computer Vision Application
Jul 21, 2026
Berkeley Lab deployed Meta’s SAM 3 and DINOv3 across 300 A100 GPUs to segment X-ray and neutron imagery, returning semantically labeled 3D volumes to beamline scientists in about 15 minutes.
Generative AI
Jul 23, 2026
Black Forest Labs opened FLUX 3 early access for image, video, and synchronized audio generation, plus FLUX-mimic, a robot-control variant that learns factory tasks from 30 minutes of demonstrations.
Video Segmentation
Aug 13, 2026
VOS-Agent combined SAM 3 with routing, tracking, and semantic agents for difficult video segmentation, scoring 69.82 J&F and taking first place in the ECCV 2026 LSVOS MOSEv2 challenge overall.
AI Resources
Video Benchmark
UVMulti released an underwater video benchmark with 100 sequences and 87,352 frames, providing pixel-level segmentation, image-enhancement annotations, and depth ground truth for multitask vision.
Motion Tracking
HumanTracker introduced a 153-hour humanoid motion benchmark and HumanScore, a preference-aligned metric trained on 12k motion pairs to expose contact, stability, and foot-skating failures.
Embodied Vision
H2R-Bench introduced a benchmark for turning human manipulation videos into robot-centric video, evaluating eleven world models on goals, contacts, embodiment accuracy, actions, and visual quality.
AI Events
ECCV 2026
Sweden ⏰ Sept 8–12
A premier biennial research conference for the global tech communities in Computer Vision and Machine Learning.
GAI World 2026
Boston ⏰ Sept 28-30
A high-level executive conference focused on enterprise AI adoption, measurable ROI, and real-world implementation case studies.
What’s New at BasicAI
Blog Pick
How Much Do Image Annotation Services Cost? The Global Benchmark
Read: An image annotation service buyer’s guide. Compare image annotation service prices by task and pricing model. See what changes the final quote.
15 Best Enterprise Image Annotation Service Providers
Read: Compare 15 enterprise image annotation service providers by delivery model, quality control, security, task coverage, and best-fit use case.
Social Media Highlights
LinkedIn: GenCeption repurposes video diffusion models into a unified, text-guided feed-forward vision system.
Facebook: Ground3D-LMM unifies 3D point grounding and real-world metric measurement in multi-turn dialogue.
Subscribe to receive monthly AI newsletter
