awesome-repositories.com
Blog
MCP
awesome-repositories.com

Descubre los mejores repositorios open-source con nuestra búsqueda potenciada por IA.

ExplorarBúsquedas curadasAlternativas open-sourceSoftware autohospedableBlogMapa del sitio
ProyectoAcerca deCómo clasificamosPrensaServidor MCP
Aviso legalPrivacidadTérminos
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
·
isl-org avatar

isl-org/MiDaSArchived

0
View on GitHub↗
5,411 estrellas·723 forks·Python·MIT·11 vistas

MiDaS

MiDaS es una librería de visión artificial de PyTorch y un modelo de estimación de profundidad monocular diseñado para predecir la profundidad de una escena a partir de imágenes individuales. Funciona como un predictor de profundidad de escena que calcula mapas de distancia para determinar la proximidad de los objetos a la cámara.

El proyecto permite la transferencia de profundidad zero-shot, lo que permite aplicar el modelo a nuevos conjuntos de datos o entornos sin necesidad de datos de entrenamiento adicionales. Se centra en la regresión de profundidad relativa para predecir mapas de profundidad invariantes a la escala.

La librería incluye un visualizador de profundidad en tiempo real para capturar transmisiones de cámara en vivo y mostrar los mapas de profundidad correspondientes. También proporciona capacidades de preprocesamiento de visión artificial para generar datos de escena estructurales para otros modelos de aprendizaje automático.

Features

  • Depth Estimation - Computes distance maps from single images using deep learning models to determine object proximity to the camera.
  • Monocular Depth Estimators - Computes distance maps from a single image to determine how far objects are from the camera.
  • Relative Depth Map Generators - Computes distance maps from images to determine how far objects are from the camera.
  • Relative Depth Estimators - Predicts scale-invariant depth maps by focusing on relative distances between objects rather than absolute metric values.
  • Zero-Shot Depth Transfer - Enables the application of depth estimation models to new datasets or environments without additional training data.
  • PyTorch Implementations - Collection of neural network implementations for depth perception built using the PyTorch framework.
  • Computer Vision Preprocessing - Generates depth maps from images to provide structural scene data for other machine learning models.
  • Mixed-Dataset Training - Combines diverse depth datasets during training to improve the model's ability to generalise across real-world environments.
  • CNN Feature Extractors - Uses a deep convolutional neural network to identify visual patterns and spatial relationships for feature extraction.
  • Live Depth Visualizations - Captures real-time camera imagery and displays corresponding depth maps for immediate visual feedback.
  • Real-time Depth Streaming - Processes live camera feeds to generate and display depth information for immediate visual feedback.
  • Depth Visualizers - Captures live camera feeds and displays corresponding depth maps for immediate visual feedback.
  • Real-Time Model Inference on Frames - Executes model forward passes on incoming camera frames to generate a continuous stream of depth maps.

Historial de estrellas

Gráfico del historial de estrellas de isl-org/midasGráfico del historial de estrellas de isl-org/midas

Búsqueda con IA

Explora más repositorios increíbles

Describe lo que necesitas en lenguaje sencillo: la IA clasifica miles de proyectos open-source curados por relevancia.

Start searching with AI

Preguntas frecuentes

¿Qué hace isl-org/midas?

MiDaS es una librería de visión artificial de PyTorch y un modelo de estimación de profundidad monocular diseñado para predecir la profundidad de una escena a partir de imágenes individuales. Funciona como un predictor de profundidad de escena que calcula mapas de distancia para determinar la proximidad de los objetos a la cámara.

¿Cuáles son las características principales de isl-org/midas?

Las características principales de isl-org/midas son: Depth Estimation, Monocular Depth Estimators, Relative Depth Map Generators, Relative Depth Estimators, Zero-Shot Depth Transfer, PyTorch Implementations, Computer Vision Preprocessing, Mixed-Dataset Training.

¿Qué alternativas de código abierto existen para isl-org/midas?

Las alternativas de código abierto para isl-org/midas incluyen: liheyoung/depth-anything — Depth-Anything is a monocular depth estimation foundation model that produces dense per-pixel depth maps from a single… depthanything/depth-anything-v2 — Depth-Anything-V2 is a computer vision foundation model designed for general-purpose spatial understanding and depth… bytedance-seed/depth-anything-3 — Depth-Anything-3 is a collection of core model implementations for depth prediction, multi-view geometry estimation,… vt-vl-lab/3d-photo-inpainting — This project is an RGB-D image inpainting tool and framework for 3D photo reconstruction. It transforms single 2D… realsenseai/librealsense — The Intel RealSense SDK is a software development kit providing drivers and libraries for interfacing with depth… nianticlabs/monodepth2 — This project is a computer vision system for monocular depth estimation and 3D point cloud generation. It provides a…

Alternativas open-source a MiDaS

Proyectos open-source similares, clasificados según cuántas características comparten con MiDaS.
  • liheyoung/depth-anythingAvatar de LiheYoung

    LiheYoung/Depth-Anything

    8,124Ver en GitHub↗

    Depth-Anything is a monocular depth estimation foundation model that produces dense per-pixel depth maps from a single RGB image. It is built on a DINOv2 Vision Transformer encoder backbone and trained on 62 million unlabeled images using a teacher-student pseudo-labeling framework, enabling robust generalization across diverse scenes without task-specific training. The model outputs both relative depth maps, which capture the ordering of scene points, and metric depth maps with real-world units after fine-tuning on datasets like NYUv2 or KITTI. The project distinguishes itself through its ab

    Pythondepth-estimationimage-synthesismetric-depth-estimation
    Ver en GitHub↗8,124
  • bytedance-seed/depth-anything-3Avatar de ByteDance-Seed

    ByteDance-Seed/Depth-Anything-3

    4,412Ver en GitHub↗

    Depth-Anything-3 is a collection of core model implementations for depth prediction, multi-view geometry estimation, and RGB-D spatial pipelines. It includes a monocular depth estimation model for predicting depth maps from single images or video, and a 3D Gaussian splatting generator that predicts parameters to synthesize high-fidelity novel views of a scene. The project provides a multi-view geometry estimator for calculating spatially consistent depth and camera poses across synchronized visual inputs. It also functions as a visual SLAM enhancement tool designed to reduce drift and improve

    Python
    Ver en GitHub↗4,412
  • depthanything/depth-anything-v2Avatar de DepthAnything

    DepthAnything/Depth-Anything-V2

    8,320Ver en GitHub↗

    Depth-Anything-V2 is a computer vision foundation model designed for general-purpose spatial understanding and depth perception. It functions as a monocular depth estimation model that predicts relative and absolute depth maps from single images or video sequences. The project provides specialized tools for both relative depth estimation and metric depth calculation, allowing for the determination of absolute physical distances in indoor and outdoor environments. It includes a video depth estimation framework that ensures temporal consistency across sequential frames to maintain stable depth

    Pythonmonocular-depth-estimation
    Ver en GitHub↗8,320
  • vt-vl-lab/3d-photo-inpaintingAvatar de vt-vl-lab

    vt-vl-lab/3d-photo-inpainting

    7,081Ver en GitHub↗

    This project is an RGB-D image inpainting tool and framework for 3D photo reconstruction. It transforms single 2D images into 3D content by estimating monocular depth and synthesizing missing color and depth data to fill occluded regions. The system uses a layered depth image representation to manage scene boundaries and pixel connectivity. This allows for novel view synthesis, enabling the generation of videos that simulate motion parallax effects from different camera perspectives. The project covers a range of spatial modeling capabilities, including depth map estimation, disparity-based

    Python
    Ver en GitHub↗7,081
  • Ver las 30 alternativas a MiDaS→