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

By Baoyuan Wu, Weidong Chen, Yanbo Fan, Yong Zhang, Jinlong Hou, T. Zhang. 2019

95 citations recorded by the source corpus.

In existing visual representation learning tasks, deep convolutional neural networks (CNNs) are often trained on images annotated with single tag, such as ImageNet. However, single tag annotation 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 codes have been released at https://github.com/Tencent/tencent-ml-images. It is expected to promote other vision tasks in the research and industry community.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

Measuring Robustness to Natural Distribution Shifts in Image Classification (2020, neutral): “…includes models trained on (i) Facebook’s collection of 1 billion Instagram images [56, 104], (ii) the YFCC 100 million dataset [104], (iii) Google’s JFT 300 million dataset [82, 102], (iv) a subset of OpenImages [98], or (v) a subset of the full ImageNet dataset of 21,841 classes [11, 49, 99].”
Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language (2022, methodology): “For Internet data, we use Tencent ML-Images (Wu et al. 2019) for object entities.”
Does Object Recognition Work for Everyone? (2019, methodology): “In addition to the cloudbased systems, we also analyzed a state-of-the-art object recognition system that was trained exclusively on publicly available data: namely, a ResNet-101 model [19] that was trained on the Tencent ML Images dataset [46] and achieves an ImageNet validation accuracy of 78.”
Multimodal datasets: misogyny, pornography, and malignant stereotypes (2021, neutral): “This is highlighted in the emergence of bigger datasets such as Tencent ML-images dataset [65] (in February 2020) that encompasses most of these non-imageable classes 30 , the continued availability of models trained on the full-ImageNet-21k dataset in repositories such as TF-hub 31 , the continued…”
Region Graph Embedding Network for Zero-Shot Learning (2020, background): “, the CNN models [46, 68, 70, 75] usually suffer from insufficient (or no) training data.”
The Emerging Trends of Multi-Label Learning (2020, background): “, Open Images [27], newly released Tencent ML-Images [28] promote deep learning for MLC problem.”
Open X-Embodiment: Robotic Learning Datasets and RT-X Models : Open X-Embodiment Collaboration0 (2023, unknown): “Even the largest data collection efforts still end up with datasets that are a fraction of the size and diversity of benchmark datasets in vision (5-18M) [4, 5] and NLP (1.5B-4.”
Sparse Adversarial Attack via Perturbation Factorization (2020, background): “Deep neural networks (DNNs) have achieved a great success in many applications, such as image classification [16,32,35], face recognition [31,34], natural”
A Framework for Deprecating Datasets: Standardizing Documentation, Identification, and Communication (2021, background): “Yet while ImageNet and Tiny Images have been remediated or withdrawn, other datasets based on them, such as CIFAR-10 and CIFAR-100 [51] and Tencent ML-Images [101] remain active.”
Do We Train on Test Data? Purging CIFAR of Near-Duplicates (2019, background): “The world wide web has become a very affordable resource for harvesting such large datasets in an automated or semiautomated manner [4, 11, 9, 20].”
Decoupled Box Proposal and Featurization with Ultrafine-Grained Semantic Labels Improve Image Captioning and Visual Question Answering (2019, background): “Prior work demonstrated the power of pre-training with image classiﬁcation at scale (Sun et al., 2017; Mahajan et al., 2018; Wu et al., 2019).”
Exact Adversarial Attack to Image Captioning via Structured Output Learning With Latent Variables (2019, background): “, image classification [15, 33, 32, 31], object detection [9], or object tracking [38, 19]), of which the loss functions are factorized to independent (i.”