Extract normalized grounding coordinates and visualizations from images and videos.
npx skills add https://github.com/zai-org/GLM-skills --skill glmv-grounding
Please help me install this Agent Skill. Skill: glmv-grounding Source: https://github.com/zai-org/GLM-skills/tree/main/skills/glmv-grounding Command: npx skills add https://github.com/zai-org/GLM-skills --skill glmv-grounding
Extracts prompt-specified target locations from images and videos and converts model outputs into reliable, normalized coordinates and visual overlays so users can automatically localize, annotate, and inspect visual targets without manual pixel math.
Run the grounding CLI with an image URL and a concise prompt to receive normalized 0-1000 coordinates and optional visualizations.
High-intent search queries and answers about installing and using this skill.
To extract bounding boxes for specific objects in images, this Skill parses GLM-V grounding outputs into standard formats like 2D boxes, 2D points, and polygons normalized to a 0-1000 coordinate range. It provides parsing, validation, and Python-based visualization utilities for object detection extraction.
To track and visualize an object across multiple video frames, this Skill supports video object tracking extraction and per-second MOT visualization. It parses video tracking JSON and generates annotated video outputs with configurable labeling, colors, and thickness for review.
Yes, you need a ZHIPU_API_KEY to run visual grounding tasks. This Skill requires GLM-V API access via this environment-based configuration to process user-specified targets and return normalized coordinates.
Yes, you can overlay bounding boxes and labels directly onto images and videos. This Skill provides Python-based visualization utilities to draw results with configurable labeling, colors, and thickness, and reverse-normalizes coordinates to pixel space for downstream use.
The coordinate format returned by GLM-V for object detection is normalized to a 0-1000 range. This Skill parses these standard formats, including 2D boxes, 3D boxes, and polygons, and can reverse-normalize them to pixel space for downstream use.
ViewWayExtract and visualize GLM-V grounding results from images or videos.
ghjghjghkimoOverlay VLM bounding box annotations from JSON onto images using Python and Pillow.
bettercallfanMatch detections or tracks against predefined camera ROI polygons.
zai-orgGenerate captions and summaries for images, videos, and documents via the GLM-V API.
