Command Palette
Search for a command to run...
Iandola Forrest N. Shen Anting Gao Peter Keutzer Kurt
Abstract
Recently, there has been a flurry of industrial activity around logorecognition, such as Ditto's service for marketers to track their brands inuser-generated images, and LogoGrab's mobile app platform for logo recognition.However, relatively little academic or open-source logo recognition progresshas been made in the last four years. Meanwhile, deep convolutional neuralnetworks (DCNNs) have revolutionized a broad range of object recognitionapplications. In this work, we apply DCNNs to logo recognition. We proposeseveral DCNN architectures, with which we surpass published state-of-artaccuracy on a popular logo recognition dataset.
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| image-classification-on-flickrlogos-32 | DeepLogo (GoogLeNet-GP) | Accuracy: 89.6 |
| object-detection-on-flickrlogos-32 | DeepLogo (VGG) | MAP: 74.4 |
| object-detection-on-flickrlogos-32 | DeepLogo (AlexNet) | MAP: 73.5 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.