AIbase
EN
Home
Oxford Pet Segmentation
O

Oxford Pet Segmentation

Developed by marinaking1
A PyTorch-based FPN architecture image segmentation model supporting multiple encoders, suitable for semantic segmentation tasks.
Image Segmentation
Safetensors
Open Source License:MIT#Image Segmentation#ResNet Encoder#High Precision IOU
Downloads 60
Release Time : 4/9/2025

Model Overview

This model adopts the Feature Pyramid Network architecture, combined with deep encoders (e.g., ResNet) to achieve efficient image semantic segmentation, supporting custom parameter configurations and pre-trained weight loading.

Model Features

Flexible Encoder Selection
Supports various pre-trained encoders (e.g., ResNet34) to adapt to different computational resource requirements.
Pyramid Feature Fusion
Achieves multi-scale feature fusion through the FPN architecture, improving segmentation accuracy.
Ready-to-Use Pre-trained Models
Provides ImageNet pre-trained model weights, supporting quick fine-tuning.

Model Capabilities

Image Semantic Segmentation
Multi-class Pixel-level Classification
Supports Custom Input Resolution

Use Cases

Pet Image Analysis
Pet Image Segmentation
Achieves precise segmentation of pets and backgrounds on the Oxford Pet dataset.
Test set IoU reaches 0.915
Medical Imaging
Organ Segmentation
Can be transferred and applied to organ recognition in CT/MRI images (requires fine-tuning).
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase