Description:

The dataset comprises over 1000+ original Television/TV images, captured and crowdsourced from over 400+ urban and rural areas. Each image has been manually reviewed and verified by computer vision professionals at DataCluster Labs, ensuring high quality and accuracy. This dataset presents a significant challenge for machine learning models due to its diverse conditions and settings.

Download Dataset

Dataset Features

Size: 1000+ images

Contributors: Over 1000+ crowdsource contributors

Resolution: 98% of the images are HD and above (1920×1080 and higher)

Geographical Diversity: Captured across 400+ cities throughout India

Lighting Conditions: Includes various lighting scenarios such as day and night, and different distances and viewpoints

Device Used: Captured using mobile phones during 2022-2023

Usage: Ideal for enhancing real-world applications such as smart home systems, retail analytics, and security monitoring

Available Annotation Formats

COCO

YOLO

PASCAL-VOC

Tf-Record

Ownership and Licensing

DataCluster Labs exclusively owns the images in this dataset, which were not sourced from the internet. A license can be purchased to access a larger portion of the training dataset for research and commercial purposes.

Applications and Use Cases

Smart Home Systems: Improve the accuracy of TV detection and interaction in smart home environments.

Retail Analytics: Utilize the dataset to analyze consumer behavior and optimize in-store experiences.

Security Monitoring: Enhance security systems by improving the detection and recognition of TVs in various settings.

Computer Vision Research: Provide a challenging and diverse dataset for developing advanced computer vision algorithms.

This dataset is sourced from Kaggle.