WXinlong/SOLO Alternatives

Choose when you need a modular detection/segmentation toolbox supporting many models rather than a single-algorithm repo.

OpenMMLab Detection Toolbox and Benchmark

Choose when you're tied to the PaddlePaddle ecosystem or need its deployment tooling for instance segmentation.

Object Detection toolkit based on PaddlePaddle. It supports object detection, instance segmentation, multiple object tracking and real-time multi-person keypoint detection.

Choose for transformer-based universal segmentation across instance, semantic, and panoptic tasks, though it is archived.

Code release for "Masked-attention Mask Transformer for Universal Image Segmentation"

Choose for a unified transformer framework that handles detection and segmentation jointly.

[CVPR 2023] Official implementation of the paper "Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation"

Choose for a modern real-time DETR-based segmentation model with active maintenance and fine-tuning support.

RF-DETR is a real-time object detection and segmentation model architecture developed by Roboflow, SOTA on COCO, designed for fine-tuning. [ICLR 2026]

Choose for a lightweight real-time instance segmentation model that predicts prototypes and coefficients when speed matters more than mask quality.

A simple, fully convolutional model for real-time instance segmentation.

Choose when you want a framework-agnostic wrapper to train instance segmentation models with Fastai or PyTorch Lightning.

An Agnostic Computer Vision Framework - Pluggable to any Training Library: Fastai, Pytorch-Lightning with more to come

Choose when you want a toolbox that includes SOLO alongside other instance-level detection models.

AdelaiDet is an open source toolbox for multiple instance-level detection and recognition tasks.

Choose for high-quality mask refinement on top of Mask R-CNN when boundary accuracy is critical.

Mask Transfiner for High-Quality Instance Segmentation, CVPR 2022

Choose for occlusion-aware instance segmentation using bilayers when handling overlapping objects.

Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers [CVPR 2021]

Choose for a legacy framework implementation of instance segmentation if you're on older infrastructure.

A Simple and Versatile Framework for Object Detection and Instance Recognition

Choose for a TensorFlow 1.x Mask R-CNN implementation if you must stay in that ecosystem.

Mask RCNN in TensorFlow

Choose for a two-stage Mask R-CNN variant with mask scoring if you need that specific accuracy metric.

Codes for paper "Mask Scoring R-CNN".

Choose for an MXNet Mask R-CNN implementation if you're on Apache MXNet.

An MXNet implementation of Mask R-CNN

Choose for a two-stage instance segmentation model with path aggregation if you need that architecture.

PANet for Instance Segmentation and Object Detection

Choose for an older anchor-free real-time instance segmentation model if you need a legacy baseline.

[CVPR 2020] CenterMask : Real-Time Anchor-Free Instance Segmentation

Choose for the improved CenterMask2 anchor-free instance segmentation, though it is no longer maintained.

[CVPR 2020] CenterMask : Real-time Anchor-Free Instance Segmentation