Idly-Dosa-Vada Dataset

This archive contains images and labels for the Idly-Dosa-Vada (IDV) dataset, for use with Yolo (and Tensorflow) object detection frameworks.

The dataset contains 1009 images, and corresponding labels.

The dataset was created by using euclidaug, using only 6 images per class.

Folder structure after extracting idv-dataset-files.zip:

out_images - contains all training images

out_labels - contains labels for each image, in Yolo format

For usage, refer to the IEEE-DL-TAP instructions, which are derived from https://github.com/prabindh/yolo-bins/tree/master/capacito

Step 1: Generate full list of image files, for use in the training process. In Windows, this is done using the below command:

dir /s/b *.jpg > trainingfile.txt

Step 2: Using the above file, and the labelled images and labels, start the training process with Yolo using instructions at https://github.com/prabindh/yolo-bins/tree/master/capacito

Step 3: Perform inference using Yolo