goyaljai/IPL-Player-Detection-IITB-PML

IPL Player Detection Dataset — IITB PML Sem1

IPL cricket player detection dataset with cell-level team annotations and player counts. Created at IIT Bombay for the Python for Machine Learning (PML) Sem 1 project.

Dataset Overview

• 1005 images from IPL broadcast footage (800×600px)
• 8×8 grid annotation per image — each of 64 cells labeled with the IPL team present
• Player count per image (0–20)
• Train/Test split: 793 train / 212 test
• 10 IPL teams: CSK, DC, GT, KKR, LSG, MI, PBKS, RR, RCB, SRH

Use Cases

• IPL team detection from broadcast frames
• Cricket player localization and counting
• Sports image segmentation
• Multi-label classification on cricket images
• Player density estimation in cricket stadiums

Label Schema

annotations.csv columns:

Column	Description
Image File Name	img_NNN.jpg
Train Or Test	Train or Test
count	Total player count in image (0–20)
c01–c64	Team ID for each of 64 grid cells (row-major, 8 cols/row)

Team IDs: 0=empty, 1=CSK, 2=DC, 3=GT, 4=KKR, 5=LSG, 6=MI, 7=PBKS, 8=RR, 9=RCB, 10=SRH

Folder Structure

train/   — 793 images (img_*.jpg)
test/    — 212 images (img_*.jpg)
annotations.csv — labels for all annotated images

Quick Start

import kagglehub
import pandas as pd
from pathlib import Path
from PIL import Image

# Download dataset
path = kagglehub.dataset_download("goyaljai0207/ipl-player-detection-iitb-pml")

# Load annotations
df = pd.read_csv(f"{path}/annotations.csv")
print(df.head())

# Load an image
img = Image.open(f"{path}/train/img_1.jpg")
img.show()

# Get label grid for first image
row = df.iloc[0]
grid = [[row[f'c{r*8+c+1:02d}'] for c in range(8)] for r in range(8)]
print(grid)

Also available on HuggingFace

from datasets import load_dataset
ds = load_dataset("goyaljai/IPL-Player-Detection-IITB-PML")

Team Distribution (1005 images, any-cell presence)

Team	Images
MI	177
RCB	153
GT	131
RR	131
CSK	130
PBKS	127
LSG	115
KKR	112
DC	110
SRH	107

Citation

If you use this dataset, please cite:

@dataset{ipl_player_detection_2026,
  title     = {IPL Player Detection Dataset},
  author    = {Goyal, Jai and contributors},
  year      = {2026},
  publisher = {Kaggle},
  url       = {https://www.kaggle.com/datasets/goyaljai0207/ipl-player-detection-iitb-pml}
}

Keywords

IPL dataset, cricket player detection, IPL team classification, cricket image dataset, broadcast frame annotation, player count dataset, cricket computer vision, IPL 2024 dataset, sports detection dataset, IITB machine learning dataset, cricket jersey detection, multi-label cricket dataset