Used the QuickDraw Dataset found on the official website. https://quickdraw.withgoogle.com/data
Tried Basic CNN, Pretrained Resnet and VGG Model. Got Accuracy of 92,94%
Used https://deepai.org/api to generate the image using the result of the image predicted.
Added the implementation of : Generative Adversarial Text to Image Synthesis : https://arxiv.org/abs/1605.05396 in the code-repo for creating own text-to-image model here
To-Do : Implement StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networks, https://arxiv.org/abs/1612.03242 for text to image generation.
- Draw on the Canvas Provided.
- Click on Submit.
- On Submit api call to model --> predicted text --> text goes to deepai api --> result shown as predicted image.
- Install the requirements.txt
- On terminal, set the Flask env : export flask_app = server_app
- then Flask run