Multi-domain Image Characteristics Dataset

The Multi-domain Image Characteristic Dataset consists of thousands of images sourced from the internet. Each image falls under one of three domains - animals, birds, or furniture. There are five types under each domain. There are 200 images of each type, summing up the total dataset to 3,000 images. The master file consists of two columns; the image name and the visible characteristics of that image. Every image was manually analyzed and the characteristics for each image were generated, ensuring accuracy.

Images falling under the same domain have a similar set of characteristics. For example, pictures under the bird's domain will have a common set of characteristics such as the color of the bird, the presence of a beak, wing, eye, legs, etc. Care has been taken to ensure that each image is as unique as possible by including pictures that have different combinations of visible characteristics present. This includes pictures having variations in the capture angle, etc.