Tencent ML-Images: A Large-Scale Multi-Label Image Database for Visual Representation Learning

Abstract

In existing visual representation learning tasks, deep convolutional neural networks (CNNs) are often trained on images annotated with single tags, such as ImageNet. However, a single tag cannot describe all important contents of one image, and some useful visual information may be wasted during training. In this work, we propose to train CNNs from images annotated with multiple tags, to enhance the quality of visual representation of the trained CNN model. To this end, we build a large-scale multi-label image database with 18M images and 11K categories, dubbed Tencent ML-Images. We efficiently train the ResNet-101 model with multi-label outputs on Tencent ML-Images, taking 90 hours for 60 epochs, based on a large-scale distributed deep learning framework,i.e.,TFplus. The good quality of the visual representation of the Tencent ML-Images checkpoint is verified through three transfer learning tasks, including single-label image classification on ImageNet and Caltech-256, object detection on PASCAL VOC 2007, and semantic segmentation on PASCAL VOC 2012. The Tencent ML-Images database, the checkpoints of ResNet-101, and all the training codehave been released at https://github.com/Tencent/tencent-ml-images. It is expected to promote other vision tasks in the research and industry community.

arXiv commentThis work is accepted to IEEE Access

References in corpus (4)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale
• WebVision Database: Visual Learning and Understanding from Web Data

Cited by in corpus (25)

• The Emerging Trends of Multi-Label Learning
• Scaling Out-of-Distribution Detection for Real-World Settings
• Measuring Robustness to Natural Distribution Shifts in Image Classification
• Multimodal datasets: misogyny, pornography, and malignant stereotypes
• Does Object Recognition Work for Everyone?
• Sewer-ML: A Multi-Label Sewer Defect Classification Dataset and Benchmark
• Do We Train on Test Data? Purging CIFAR of Near-Duplicates
• MotionBenchMaker: A Tool to Generate and Benchmark Motion Planning Datasets
• REFLACX, a dataset of reports and eye-tracking data for localization of abnormalities in chest x-rays
• Large image datasets: A pyrrhic win for computer vision?
• The effective noise of Stochastic Gradient Descent
• Semantic Redundancies in Image-Classification Datasets: The 10% You Don't Need
• Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts
• Webly Supervised Image Classification with Metadata: Automatic Noisy Label Correction via Visual-Semantic Graph
• sigmoidF1: A Smooth F1 Score Surrogate Loss for Multilabel Classification
• Exact Adversarial Attack to Image Captioning via Structured Output Learning with Latent Variables
• On The State of Data In Computer Vision: Human Annotations Remain Indispensable for Developing Deep Learning Models
• EvoSplit: An evolutionary approach to split a multi-label data set into disjoint subsets
• Quantized Adam with Error Feedback
• Knowledge-Based Construction of Confusion Matrices for Multi-Label Classification Algorithms using Semantic Similarity Measures
• Cross-modal supervised learning for better acoustic representations
• PCHands: PCA-based Hand Pose Synergy Representation on Manipulators with N-DoF
• Privacy for Rescue: A New Testimony Why Privacy is Vulnerable In Deep Models
• Decoupled Box Proposal and Featurization with Ultrafine-Grained Semantic Labels Improve Image Captioning and Visual Question Answering
• Multi-Label Learning from Single Positive Labels