Preparing the Dataset
As is the case with any image classifier, the model needs to be trained on a dataset from which it can infer and extract the features corresponding to a particular category. The BTS Image Classifier would contain 7 categories (total number of members). The dataset can be prepared by manually collecting images of different members and then clubbing them in a folder of that category. In order to fasten this process, a Python script can be employed to create the dataset. The script would fetch images from Google Image Search. (Disclaimer: Using these images may lead to a copyright violation so proceed at your own risk).
A folder called simple_images would appear at the location where the script is present. Inside the simple_images folder, the folders corresponding to each of the seven members with 150 images would be present.
It is time to code the classifier. It is recommended to use Google Collab (the GPU would come in handy while training) and have the dataset uploaded to Google Drive.
Python3
# Import fastbook from fastbook import *from fastai.vision.widgets import *from google.colab import drive drive.mount('/content/drive') import fastbook fastbook.setup_book() class DataLoaders(GetAttr): def __init__(self, *loaders): self.loaders = loaders def __getitem__(self, i): return self.loaders[i] train, valid = add_props(lambda i, self: self[i]) |
DataLoaders is a class that is responsible for providing the valid and train dataset to the model.
Python3
# Import the required function to download from the Simple Image Download library. from simple_image_download import simple_image_download as simp # Create a response instance. response = simp.simple_image_download # The following lines would look up Google Images and download the number of images specified. # The first argument is the term to search, and the second argument is the number of images to be downloaded. response.download('BTS Jin', 150) response.download('BTS Jimin', 150) response.download('BTS RM', 150) response.download('BTS J-Hope', 150) response.download('BTS Suga', 150) response.download('BTS Jungkook', 150) |
Identify Members of BTS — An Image Classifier
BTS is an eminent K-Pop band comprising of 7 members. This article looks at an image classifier that would recognize the name of the band member from a picture. The image classifier would be built using fastai. It is a deep learning library that aims to democratize deep learning. It is built on top of PyTorch, and has plethora of models with optimized weights that are ready-to-use. The application would be hosted on Binder, and the end product would look like this:
Contact Us