AIbase
EN
Home
Segformer B1 Finetuned Tennisdata
S

Segformer B1 Finetuned Tennisdata

Developed by julia-wenkmann
An image segmentation model based on the nvidia/mit-b0 architecture, optimized for tennis scenes to identify key elements such as players and courts
Image Segmentation
Transformers
Open Source License:Other#Tennis scene segmentation#Player positioning and tracking#Lightweight image segmentation
Downloads 34
Release Time : 2/20/2024

Model Overview

This model is a lightweight version (B1) of the SegFormer architecture, fine-tuned on tennis match datasets, suitable for image segmentation tasks in tennis match scenarios

Model Features

Tennis scene optimization
Fine-tuned specifically for tennis match scenes, accurately identifying key elements such as players and courts
Lightweight architecture
Utilizes the MIT-B0 lightweight backbone network, reducing computational resource requirements while maintaining performance
Multi-category segmentation
Supports precise segmentation of multiple categories including top players, bottom players, and courts

Model Capabilities

Image segmentation
Scene understanding
Sports analysis

Use Cases

Sports analysis
Tennis match video analysis
Automatically segments player and court regions in match videos
Top player recognition accuracy 70.71%, bottom player 89.04%
Player trajectory tracking
Provides foundational segmentation data for player movement analysis
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