For a very long period, object detection in visual platforms has been researched. It can be used to detect people walking, cars driving, classify people by age or gender, and more. Finding a target's exact placement within a scene is the basic goal of object detection. Features must be robust, differential, and simple to calculate in order to detect objects accurately in real-time. Social security benefits from the categorization of children and adults since it makes it much simpler to apprehend criminals based on bodily characteristics. Several research of classification between children and adults have been introduced for categorization. In light of it, the aim of this project is to divide pedestrians into two groups, namely children and adults.
The dataset contains 2 folders: one with test data and the other one with train data. The train-test-split ratio is 0.15, with the test dataset containing 120 images and the train dataset containing 680. The images have a resolution of 370x320 pixels in RGB color model. Both the folders contain 2 classes:
Adults
Children
The dataset was obtained downloading images from internet websites such as https://www.istockphoto.com and https://pixabay.com. The obtained images were randomly shuffled and resized so that all the images had a resolution of 370x320 pixels. Then, they were split into train and test datasets and saved.