Rakuten Deep Learning Challenge
Our Bengaluru office just launched a deep learning challenge competition running until the end of November, and we are looking for as many participants as possible to join. We already have over 500 people signed up, but the more the merrier. There are also some pretty sizable prizes available, with the winner taking home 200,000 Indian Rupees.
Advancement in image processing and machine learning algorithms have allowed us to accurately detect and classify objects with high degree of precision. Can this technology be extended to images related to food recipes? In Rakuten Deep Learning Challenge, we are asking you to classify food images into different categories. This will enable better taxonomy and improved indexing for search operations. We bring you hundreds of thousands of images from Rakuten Recipe – Rakuten’s haven for food lovers, home chefs and really just anyone with an empty stomach and an appetite for new flavors.
A labelled dataset is provided for supervised training and mapping to different food categories. The Rakuten recipe food image data is distributed across 43 categories with around 0.79 million images. Each image category is prepared according to their Recipe, time taken to prepare it and its ingredients etc. The challenge is to develop an algorithm that would classify a given recipe food image into one of the pre-defined categories with high precision and recall.
Are you up for the challenge? If so, make sure to see the official page linked below for more details and how to sign up!