Tom And Jerry Image Classification
Collection of 5k+ images with labelled data of Tom & Jerry cartoon show
@kaggle.balabaskar_tom_and_jerry_image_classification
Collection of 5k+ images with labelled data of Tom & Jerry cartoon show
@kaggle.balabaskar_tom_and_jerry_image_classification
Tom and Jerry is an American animated media franchise and series of comedy short films created in 1940 by William Hanna and Joseph Barbera. Best known for its 161 theatrical short films by Metro-Goldwyn-Mayer, the series centers on the rivalry between the titular characters of a cat named Tom and a mouse named Jerry.
This is one of the famous cartoon shows, that we would have never missed watching during our childhood. Now, its time to use our deep learning skills to detect our favorite characters - Tom and Jerry in the images extracted from some of the shows.
This dataset contains more than 5k images (exactly 5478 images) extracted from some of Tom & Jerry's show videos, that are available online.
The downloaded videos are converted into images with 1 frame per second (1 FPS)
Labeled images are separated into 4 different folders as given.
Folder - tom_and_jerry
SubFolder tom - contains images only with 'tom'
SubFolder jerry - contains images only with 'jerry'
SubFolder tom_jerry_1 - contains images with both 'tom' and 'jerry'
SubFolder tom_jerry_0 - contains images without both the characters
There are images that can be challenged during training in image classification, as these images are distorted in the original size or shape and color of the characters. These image details are given in csv file - challenges.csv Doing error analysis on these images after model training, will help us understand how to improve the score.
Few examples are given below:
Anyone who has the link will be able to view this.