Tag
A hands-on benchmark of TypeSafe's Jev 1.14 image input and new Box (bounding-box) output across 48 labeled COCO photos, finding strong yes/no accuracy (98.4%) but notable pitfalls: Box never answers "not found", confidence thresholds don't transfer between model versions, and wording strongly affects probabilities. The author recommends pairing every Box with a presence check, per-question thresholds, and pinning the model version.
Jev decision model now supports multimodal input, allowing images in both state and questions, plus a new Box type that returns YOLO-style bounding boxes for object localization. The update also supports parallel question answering for interactive vision applications, with images and Box still in experimental status.
GroundingPI 与 XPeng、北大、港大合作发布了3个开源模型：4B基础模型（检测、指代、OCR等多能力）、基于扩散并行解码的快速grounding模型GroundAnything，以及自回归VLM版本的高精度grounding模型。
A developer shares their achievement of reducing object detection latency to below 0.35 seconds on local hardware, highlighting progress in AI performance optimization.
PAANI is an on-device perception-to-guidance architecture for river-robot simulation that combines YOLO11n and MobileNetV3-Small models with evidence fusion on Arduino UNO Q to provide explainable advisories for navigation, evaluated with promising results for edge AI applications.
This paper proposes an alternative signal-centric method for processing sonar data in remote sensing applications, using CSV format and acoustic processing to reduce processing time by 91.18% and improve machine learning-driven object detection.
The user explores adding AI object detection to older security cameras without straining network bandwidth, considering edge processing with on-premises appliances to analyze video locally and only send alerts or short clips, while seeking advice on hybrid architecture and API integration.
The paper introduces TAPe+MLv3, a compact computer vision system using structured representation for multi-task tasks, achieving competitive performance on benchmarks like COCO with fewer than 100,000 parameters.
Adversarial fashion is emerging as a creative response to AI-powered surveillance, using colorful patterns and designs to evade object detection systems, with projects and products like noRecognition and Cap_able already available.
A demonstration highlights how drone footage of crowded streets can be analyzed with just 10 lines of Python to detect and count hundreds of people, illustrating how accessible computer vision has become.
This paper introduces Spatially-Sparse Linear Attention for low-latency object detection with event cameras, achieving state-of-the-art accuracy with 20x less computation than prior asynchronous methods.
Structured Prior Knowledge (SPK) is a framework that explicitly extracts latent semantic, geometric, and contextual priors from pretrained object detectors to achieve state-of-the-art out-of-distribution detection, improving interpretability and reliability.
OpenAI's GPT-5.6 Sol is benchmarked as their best vision model yet, showing significant improvements in object detection and other visual tasks compared to previous models like GPT-5.5.
Tencent researchers propose YOLO-PEFT, a structure-aware framework that treats adapter placement as constraint planning, outperforming full fine-tuning on YOLO detectors while using significantly fewer trainable parameters.
VLX-Seek-1.5-10B is an open-source 10B vision-language model from omlab, designed for fine-grained visual grounding in embodied scenarios like drones, robots, and surveillance, using region-reference localization instead of coordinate generation.
YOLO-PEFT is a structure-aware framework that formulates adapter placement as constraint planning for parameter-efficient fine-tuning of YOLO detectors, achieving better mAP than full fine-tuning with reduced memory.
SkalskiP highlights Qwen3.8-Max, a vision-language model for object detection that can be prompted with positive and negative boxes to generate detections, achieving 60-80% mAP with single or multiple prompts and performing well on diverse image types.
A tweet claims Qwen3.8-Max is the best object detection VLM, excelling across satellite, infrared, document, and hand-drawn images, with examples shared.
This paper introduces SULAND v2, a refined RGB surface landmine detection dataset and benchmark for UAV/UGV-based surveys, addressing annotation errors and domain-shift evaluation in object detection.
A tutorial on building a real-time object detection and tracking pipeline for robotics using ROS 2 and YOLOv11, covering threaded inference, ByteTrack integration, confidence validation, and ONNX export for edge deployment.