OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Evaluation & Observability

Map

Map is a Python command-line tool that evaluates object-recognition neural networks by comparing per-image detection results with labeled ground truth and calculating mean Average Precision.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Map — OpenViST project profile in AI Evaluation & Observability.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Map when you already have object-detection predictions and labeled ground-truth files and want to evaluate them with its documented mAP procedure.

Who it is for

  • Computer-vision practitioners evaluating object-detection model outputs.
  • Teams comparing model performance against a consistent labeled dataset.

Documented uses

  • Calculate per-class Average Precision and an overall mean Average Precision score.
  • Compare confidence-ranked detections with ground-truth bounding boxes.
  • Track mAP changes between model versions using consistent evaluation inputs.
Before you choose

Requirements & limits

The public GitHub repository identifies the project as open source under the Apache-2.0 license.

  • Matching uses the same class label and an IoU threshold of at least 0.5.
  • The final mAP is the mean of class-level AP scores for classes in the ground truth.
  • Python is required; Matplotlib for plots and OpenCV for animation are optional.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Calculate mAP from prepared detection files

Run the evaluator on per-image ground-truth and detection-result files in its required format.

Before you start

  • Python
  • A local clone of the Map repository
  • One ground-truth text file and one detection-results text file per image, with matching file names
  • Ground-truth rows formatted as with optional difficult marker
  • Detection rows formatted as

Steps

  1. Clone the repository from https://github.com/Cartucho/mAP.

  2. Place ground-truth files in input/ground-truth/ and detection-results files in input/detection-results/.

  3. Run python main.py.

Expected outcome

The program calculates Average Precision per ground-truth class and reports a mean Average Precision score.

Guide limitations

  • A match requires the same class label and IoU of at least 0.5; each ground-truth object can count as a true positive only once.
  • The score reflects the supplied ground truth and class set.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
Cartucho/mAP · Official homepage title · retained in its source language
How we review sourcesSuggest a correction