Comparison of State-of-the-Art Deep Learning APIs for Image Multi-Label Classification using Semantic Metrics

TL;DR

This paper evaluates and compares 13 deep learning APIs for multi-label image classification using semantic metrics to account for differences in label vocabularies, revealing insights into their true performance.

Contribution

It introduces semantic similarity metrics for evaluating image classification APIs, addressing vocabulary mismatch issues in performance assessment.

Findings

Microsoft, Imagga, and IBM APIs excel with traditional metrics.

Semantic metrics highlight InceptionResNet-v2, Inception-v3, ResNet50 as top semantic performers.

Evaluation on Visual Genome and Open Images datasets provides comprehensive performance insights.

Abstract

Image understanding heavily relies on accurate multi-label classification. In recent years, deep learning algorithms have become very successful for such tasks, and various commercial and open-source APIs have been released for public use. However, these APIs are often trained on different datasets, which, besides affecting their performance, might pose a challenge to their performance evaluation. This challenge concerns the different object-class dictionaries of the APIs' training dataset and the benchmark dataset, in which the predicted labels are semantically similar to the benchmark labels but considered different simply because they have different wording in the dictionaries. To face this challenge, we propose semantic similarity metrics to obtain richer understating of the APIs predicted labels and thus their performance. In this study, we evaluate and compare the performance of 13 of the most prominent commercial and open-source APIs in a best-of-breed challenge on the Visual Genome and Open Images benchmark datasets. Our findings demonstrate that, while using traditional metrics, the Microsoft Computer Vision, Imagga, and IBM APIs performed better than others. However, applying semantic metrics also unveil the InceptionResNet-v2, Inception-v3, and ResNet50 APIs, which are trained only with the simple ImageNet dataset, as challengers for top semantic performers.