Deep Learning Logo Detection with Data Expansion by Synthesising Context
This addresses the challenge of high labeling costs for logo detection in real-world images, offering a practical solution for applications in brand monitoring and advertising, though it is incremental in nature.
The paper tackles the problem of logo detection in unconstrained images with limited labeled data by proposing a method to synthesize training images, which significantly improves detection performance, achieving superior results on benchmark datasets like FlickrLogo-32 and TopLogo-10.
Logo detection in unconstrained images is challenging, particularly when only very sparse labelled training images are accessible due to high labelling costs. In this work, we describe a model training image synthesising method capable of improving significantly logo detection performance when only a handful of (e.g., 10) labelled training images captured in realistic context are available, avoiding extensive manual labelling costs. Specifically, we design a novel algorithm for generating Synthetic Context Logo (SCL) training images to increase model robustness against unknown background clutters, resulting in superior logo detection performance. For benchmarking model performance, we introduce a new logo detection dataset TopLogo-10 collected from top 10 most popular clothing/wearable brandname logos captured in rich visual context. Extensive comparisons show the advantages of our proposed SCL model over the state-of-the-art alternatives for logo detection using two real-world logo benchmark datasets: FlickrLogo-32 and our new TopLogo-10.