Image similarity using Deep CNN and Curriculum Learning
This work addresses image retrieval for applications like search and recommendation, but it is incremental as it builds on existing siamese and curriculum learning methods.
The paper tackled the problem of image similarity by proposing SimNet, a deep siamese network with a novel online pair mining strategy inspired by curriculum learning and a multi-scale CNN for joint embeddings, resulting in improved fine-grained similarity capture compared to traditional CNNs.
Image similarity involves fetching similar looking images given a reference image. Our solution called SimNet, is a deep siamese network which is trained on pairs of positive and negative images using a novel online pair mining strategy inspired by Curriculum learning. We also created a multi-scale CNN, where the final image embedding is a joint representation of top as well as lower layer embedding's. We go on to show that this multi-scale siamese network is better at capturing fine grained image similarities than traditional CNN's.